59 Commits

Author SHA1 Message Date
Dan Milne
abbb11a41d Return only scopes requested, add tests ( OpenID conformance test )
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / scan_container (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-02 14:55:06 +11:00
Dan Milne
b2030df8c2 Return only scopes requested ( OpenID conformance test. Update README 2026-01-02 14:05:54 +11:00
Dan Milne
07cddf5823 Version bump
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / scan_container (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-02 12:57:28 +11:00
Dan Milne
46aa983189 Don't use secret scanner for trivy - github already does it and it's hard to ignore the test key
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / scan_container (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-02 12:56:03 +11:00
Dan Milne
d0d79ee1da Try ignore capybara's test tripping trivy
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / scan_container (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-02 12:52:24 +11:00
Dan Milne
2f6a2c7406 Update ruby 3.4.6 -> 3.4.7. Update gems. Add trivy scanning and ignore unfixable Debian CVEs. Ignore a test fixture key for Capybara
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / scan_container (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-02 12:48:40 +11:00
Dan Milne
5137a25626 Add remainging rate limits. Add docker compose production example. Update beta-checklist.
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-02 12:14:13 +11:00
Dan Milne
fed7c3cedb Some beta-checklist updates 2026-01-02 11:53:41 +11:00
Dan Milne
e288fcad7c Remove old docs 2026-01-01 21:04:26 +11:00
Dan Milne
c1c6e0112e ADd backup / restore documentation
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 15:40:49 +11:00
Dan Milne
7f834fb7fa Version bump 2026-01-01 15:27:19 +11:00
Dan Milne
ae99d3d9cf Fix webauthn bug. Fix tests. Update docs
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 15:24:56 +11:00
Dan Milne
1afcd041f9 Update README, fix a test 2026-01-01 15:17:28 +11:00
Dan Milne
71198340d0 fix tests and add a Claude.md file
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 15:11:46 +11:00
Dan Milne
d597ca8810 Fix tests 2026-01-01 14:52:24 +11:00
Dan Milne
9b81aee490 Fix linting error
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 13:45:10 +11:00
Dan Milne
265518ab25 Move integration tests into right directory
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 13:43:13 +11:00
Dan Milne
adb789bbea Fix StandardRB
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 13:35:37 +11:00
Dan Milne
93a0edb0a2 StandardRB fixes
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 13:29:44 +11:00
Dan Milne
7d3af2bcec SRB fixes 2026-01-01 13:19:17 +11:00
Dan Milne
c03034c49f Add files to support brakeman and standardrb. Fix some SRB warnings 2026-01-01 13:18:30 +11:00
Dan Milne
9234904e47 Add security-todo and beta-checklists, and some security rake tasks
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2026-01-01 13:06:54 +11:00
Dan Milne
e36a9a781a Add new claims to the discovery endpoint
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 17:27:28 +11:00
Dan Milne
d036e25fef Add auth_time, acr and azp support for OIDC claims 2025-12-31 17:07:54 +11:00
Dan Milne
fcdd2b6de7 Continue adding auth_time - need it in the refresh token too, so we can accurately create new access tokens. 2025-12-31 16:57:28 +11:00
Dan Milne
3939ea773f We already have a login_time stored - the time stamp of the Session instance creation ( created after successful login ). 2025-12-31 16:45:45 +11:00
Dan Milne
4b4afe277e Include auth_time in ID token. Switch from upsert -> find_and_create_by so we actually get sid values for consent on the creation of the record 2025-12-31 16:36:32 +11:00
Dan Milne
364e6e21dd Fixes for tests and AR Encryption
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 16:08:05 +11:00
Dan Milne
9d352ab8ec Fix tests - add missing files 2025-12-31 16:01:31 +11:00
Dan Milne
d1d4ac745f Version bump 2025-12-31 15:48:52 +11:00
Dan Milne
3db466f5a2 Switch Access / Refresh tokens / Auth Code from bcrypt ( and plain ) to hmac. BCrypt is for low entropy passwords and prevents dictionary attacks - HMAC is suitable for 256-bit random data.
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 15:48:32 +11:00
Dan Milne
7c6ae7ab7e Store only HMAC'd Auth codes, rather than plain text auth codes. 2025-12-31 15:00:00 +11:00
Dan Milne
ed7ceedef5 Include the hash of the access token in the JWT / ID Token under the key at_hash as per the requirements. Update the discovery endpoint to describe subject_type as 'pairwise', rather than 'public', since we do pairwise subject ids.
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 14:45:38 +11:00
Dan Milne
40815d3576 Use SolidQueue in production. Use the find_by_token method, rather than iterating over refresh tokens, as we already fixed for tokens 2025-12-31 14:32:34 +11:00
Dan Milne
a17c08c890 Improve the README 2025-12-31 14:31:53 +11:00
Dan Milne
4f31fadc6c Improve the README and remove incorrect claims.
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 12:17:15 +11:00
Dan Milne
29c0981a59 Improve readme and tests
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 11:56:09 +11:00
Dan Milne
9d402fcd92 Clean up and secure web_authn controller
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 11:44:11 +11:00
Dan Milne
9530c8284f Version bump
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-31 10:35:27 +11:00
Dan Milne
bb5aa2e6d6 Add rails encryption for totp - allow configuration of encryption secrets from env, or derive them from SECRET_KEY_BASE. Don't leak email address via web_authn, rate limit web_authn, escape oidc state value, require password for changing email address, allow settings the hmac secret for token prefix generation 2025-12-31 10:33:56 +11:00
Dan Milne
cc7beba9de PKCE is now default enabled. You can now create public / no-secret apps OIDC apps 2025-12-31 09:22:18 +11:00
Dan Milne
00eca6d8b2 Default deny forward_auth requests 2025-12-30 16:04:01 +11:00
Dan Milne
32235f9647 version bump
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-30 11:58:31 +11:00
Dan Milne
71d59e7367 Remove plain text token from everywhere
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-30 11:58:11 +11:00
Dan Milne
99c3ac905f Add a token prefix column, generate the token_prefix and the token_digest, removing the plaintext token from use. 2025-12-30 09:45:16 +11:00
Dan Milne
0761c424c1 Fix tests. Remove tests which test rails functionality
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-30 00:18:19 +11:00
Dan Milne
2a32d75895 Fix tests - don't test standard rails features 2025-12-29 19:45:01 +11:00
Dan Milne
4c1df53fd5 Fix more tests
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-29 19:22:08 +11:00
Dan Milne
acab15ce30 Fix more tests 2025-12-29 18:48:41 +11:00
Dan Milne
0361bfe470 Fix forward_auth bugs - including disabled apps still working. Fix forward_auth tests
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-29 15:37:12 +11:00
Dan Milne
5b9d15584a Add more rate limiting, and more restrictive headers 2025-12-29 13:29:14 +11:00
Dan Milne
898fd69a5d Add permissions initializer and missing image paste controller
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-29 13:27:30 +11:00
Dan Milne
9cf01f7c7a Bump versoin 2025-12-28 14:43:26 +11:00
Dan Milne
ab362aabac Remove the rate limit for the forward auth system
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-12-28 14:40:53 +11:00
Dan Milne
283feea175 Update depenencies, bump versoin 2025-11-30 23:13:25 +11:00
Dan Milne
7af8624bf8 Handle empty backchannel logout urls
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-11-27 19:19:34 +11:00
Dan Milne
f8543f98cc Add a subdirectory for active storage
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-11-27 19:12:09 +11:00
Dan Milne
6be23c2c37 Add backchannel logout, per application logout.
Some checks failed
CI / scan_ruby (push) Has been cancelled
CI / scan_js (push) Has been cancelled
CI / lint (push) Has been cancelled
CI / test (push) Has been cancelled
CI / system-test (push) Has been cancelled
2025-11-27 16:38:27 +11:00
Dan Milne
eb2d7379bf Backchannel complete - improve oidc credential display 2025-11-27 11:52:25 +11:00
133 changed files with 9278 additions and 2183 deletions

View File

@@ -1,5 +1,21 @@
# Rails Configuration
SECRET_KEY_BASE=generate-with-bin-rails-secret
# SECRET_KEY_BASE is used for:
# - Session cookie encryption
# - Signed token verification
# - ActiveRecord encryption (currently: TOTP secrets)
# - OIDC token prefix HMAC derivation
#
# CRITICAL: Do NOT change SECRET_KEY_BASE after deployment. Changing it will:
# - Invalidate all user sessions (users must re-login)
# - Break encrypted data (users must re-setup 2FA)
# - Invalidate all OIDC access/refresh tokens (clients must re-authenticate)
#
# Optional: Override encryption keys with env vars for key rotation:
# - ACTIVE_RECORD_ENCRYPTION_PRIMARY_KEY
# - ACTIVE_RECORD_ENCRYPTION_DETERMINISTIC_KEY
# - ACTIVE_RECORD_ENCRYPTION_KEY_DERIVATION_SALT
# - OIDC_TOKEN_PREFIX_HMAC
SECRET_KEY_BASE=generate-with-bin/rails/secret
RAILS_ENV=development
# Database

View File

@@ -19,7 +19,9 @@ jobs:
bundler-cache: true
- name: Scan for common Rails security vulnerabilities using static analysis
run: bin/brakeman --no-pager
run: bin/brakeman --no-pager --no-exit-on-warn
# Note: 2 weak warnings exist and are documented as acceptable
# See docs/beta-checklist.md for details
- name: Scan for known security vulnerabilities in gems used
run: bin/bundler-audit
@@ -39,10 +41,36 @@ jobs:
- name: Scan for security vulnerabilities in JavaScript dependencies
run: bin/importmap audit
scan_container:
runs-on: ubuntu-latest
permissions:
security-events: write # Required for uploading SARIF results
contents: read
steps:
- name: Checkout code
uses: actions/checkout@v5
- name: Build Docker image
run: docker build -t clinch:${{ github.sha }} .
- name: Run Trivy vulnerability scanner
uses: aquasecurity/trivy-action@master
with:
image-ref: clinch:${{ github.sha }}
format: 'sarif'
output: 'trivy-results.sarif'
severity: 'CRITICAL,HIGH'
scanners: 'vuln' # Only scan vulnerabilities, not secrets (avoids false positives in vendored gems)
- name: Upload Trivy results to GitHub Security tab
uses: github/codeql-action/upload-sarif@v3
if: always()
with:
sarif_file: 'trivy-results.sarif'
lint:
runs-on: ubuntu-latest
env:
RUBOCOP_CACHE_ROOT: tmp/rubocop
steps:
- name: Checkout code
uses: actions/checkout@v5
@@ -52,18 +80,8 @@ jobs:
with:
bundler-cache: true
- name: Prepare RuboCop cache
uses: actions/cache@v4
env:
DEPENDENCIES_HASH: ${{ hashFiles('.ruby-version', '**/.rubocop.yml', '**/.rubocop_todo.yml', 'Gemfile.lock') }}
with:
path: ${{ env.RUBOCOP_CACHE_ROOT }}
key: rubocop-${{ runner.os }}-${{ env.DEPENDENCIES_HASH }}-${{ github.ref_name == github.event.repository.default_branch && github.run_id || 'default' }}
restore-keys: |
rubocop-${{ runner.os }}-${{ env.DEPENDENCIES_HASH }}-
- name: Lint code for consistent style
run: bin/rubocop -f github
run: bin/standardrb
test:
runs-on: ubuntu-latest

View File

@@ -1 +1 @@
3.4.6
3.4.8

7
.standard.yml Normal file
View File

@@ -0,0 +1,7 @@
ignore:
- 'test_*.rb' # Ignore test files in root directory
- 'tmp/**/*'
- 'vendor/**/*'
- 'node_modules/**/*'
- 'config/initializers/csp_local_logger.rb' # Complex CSP logger with intentional block structure
- 'config/initializers/sentry_subscriber.rb' # Sentry subscriber with module structure

48
.trivyignore Normal file
View File

@@ -0,0 +1,48 @@
# Trivy ignore file
# This file tells Trivy to skip specific vulnerabilities or files
# See: https://aquasecurity.github.io/trivy/latest/docs/configuration/filtering/
# =============================================================================
# False Positives - Test Fixtures
# =============================================================================
# Capybara test fixture - not a real private key
# Ignore secrets in test fixtures
# Format: secret:<rule-id>:<exact-file-path>
secret:private-key:/usr/local/bundle/ruby/3.4.0/gems/capybara-3.40.0/spec/fixtures/key.pem
# =============================================================================
# Unfixable CVEs - No Patches Available (Status: affected/fix_deferred)
# =============================================================================
# GnuPG vulnerabilities - not used by Clinch at runtime
# Low risk: dirmngr/gpg tools not invoked during normal operation
CVE-2025-68973
# Image processing library vulnerabilities
# Low risk for Clinch: Only admins upload images (app icons), not untrusted users
# Waiting on Debian security team to release patches
# ImageMagick - Integer overflow (32-bit only)
CVE-2025-66628
# glib - Integer overflow in URI escaping
CVE-2025-13601
# HDF5 - Critical vulnerabilities in scientific data format library
CVE-2025-2153
CVE-2025-2308
CVE-2025-2309
CVE-2025-2310
# libmatio - MATLAB file format library
CVE-2025-2338
# OpenEXR - Image format vulnerabilities
CVE-2025-12495
CVE-2025-12839
CVE-2025-12840
CVE-2025-64181
# libvips - Image processing library
CVE-2025-59933

65
Claude.md Normal file
View File

@@ -0,0 +1,65 @@
# Claude Code Guidelines for Clinch
This document provides guidelines for AI assistants (Claude, ChatGPT, etc.) working on this codebase.
## Project Context
Clinch is a lightweight identity provider (IdP) supporting:
- **OIDC/OAuth2** - Standard OAuth flows for modern apps
- **ForwardAuth** - Trusted-header SSO for reverse proxies (Traefik, Caddy, Nginx)
- **WebAuthn/Passkeys** - Passwordless authentication
- Group-based access control
Key characteristics:
- Rails 8 application with SQLite database
- Focus on simplicity and self-hosting
- No external dependencies for core functionality
## Testing Guidelines
### Do Not Test Rails Framework Functionality
When writing tests, focus on testing **our application's specific behavior and logic**, not standard Rails framework functionality.
**Examples of what NOT to test:**
- Session isolation between users (Rails handles this)
- Basic ActiveRecord associations (Rails handles this)
- Standard cookie signing/verification (Rails handles this)
- Default controller rendering behavior (Rails handles this)
- Infrastructure-level error handling (database connection failures, network issues, etc.)
**Examples of what TO test:**
- Forward auth business logic (group-based access control, header configuration, etc.)
- Custom authentication flows
- Application-specific session expiration behavior
- Domain pattern matching logic
- Custom response header generation
**Why:**
Testing Rails framework functionality adds no value and can create maintenance burden. Trust that Rails works correctly and focus tests on verifying our application's unique behavior.
### Integration Test Patterns
**Session handling:**
- Do NOT manually manipulate cookies in integration tests
- Use the session provided by the test framework
- To get the actual session ID, use `Session.last.id` after sign-in, not `cookies[:session_id]` (which is signed)
**Application setup:**
- Always create Application records for the domains you're testing
- Use wildcard patterns (e.g., `*.example.com`) when testing multiple subdomains
- Remember: `*` matches one level only (`*.example.com` matches `app.example.com` but NOT `sub.app.example.com`)
**Header assertions:**
- Always normalize header names to lowercase when asserting (HTTP headers are case-insensitive)
- Use `response.headers["x-remote-user"]` not `response.headers["X-Remote-User"]`
**Avoid threading in integration tests:**
- Rails integration tests use a single cookie jar
- Convert threaded tests to sequential requests instead
### Common Testing Pitfalls
1. **Don't test concurrent users with manual cookie manipulation** - Integration tests can't properly simulate multiple concurrent sessions
2. **Don't expect `cookies[:session_id]` to be the actual ID** - It's a signed cookie value
3. **Don't assume wildcard patterns match multiple levels** - `*.domain.com` only matches one level

View File

@@ -8,14 +8,17 @@
# For a containerized dev environment, see Dev Containers: https://guides.rubyonrails.org/getting_started_with_devcontainer.html
# Make sure RUBY_VERSION matches the Ruby version in .ruby-version
ARG RUBY_VERSION=3.4.6
ARG RUBY_VERSION=3.4.8
FROM docker.io/library/ruby:$RUBY_VERSION-slim AS base
LABEL org.opencontainers.image.source=https://github.com/dkam/clinch
# Rails app lives here
WORKDIR /rails
# Install base packages
# Install base packages and upgrade to latest security patches
RUN apt-get update -qq && \
apt-get upgrade -y && \
apt-get install --no-install-recommends -y curl libjemalloc2 libvips sqlite3 && \
ln -s /usr/lib/$(uname -m)-linux-gnu/libjemalloc.so.2 /usr/local/lib/libjemalloc.so && \
rm -rf /var/lib/apt/lists /var/cache/apt/archives

21
Gemfile
View File

@@ -35,18 +35,19 @@ gem "jwt", "~> 3.1"
gem "webauthn", "~> 3.0"
# Public Suffix List for domain parsing
gem "public_suffix", "~> 6.0"
gem "public_suffix", "~> 7.0"
# Error tracking and performance monitoring (optional, configured via SENTRY_DSN)
gem "sentry-ruby", "~> 5.18"
gem "sentry-rails", "~> 5.18"
gem "sentry-ruby", "~> 6.2"
gem "sentry-rails", "~> 6.2"
# Windows does not include zoneinfo files, so bundle the tzinfo-data gem
gem "tzinfo-data", platforms: %i[ windows jruby ]
gem "tzinfo-data", platforms: %i[windows jruby]
# Use the database-backed adapters for Rails.cache and Action Cable
gem "solid_cache"
gem "solid_cable"
gem "solid_queue", "~> 1.2"
# Reduces boot times through caching; required in config/boot.rb
gem "bootsnap", require: false
@@ -62,7 +63,7 @@ gem "image_processing", "~> 1.2"
group :development, :test do
# See https://guides.rubyonrails.org/debugging_rails_applications.html#debugging-with-the-debug-gem
gem "debug", platforms: %i[ mri windows ], require: "debug/prelude"
gem "debug", platforms: %i[mri windows], require: "debug/prelude"
# Audits gems for known security defects (use config/bundler-audit.yml to ignore issues)
gem "bundler-audit", require: false
@@ -70,8 +71,8 @@ group :development, :test do
# Static analysis for security vulnerabilities [https://brakemanscanner.org/]
gem "brakeman", require: false
# Omakase Ruby styling [https://github.com/rails/rubocop-rails-omakase/]
gem "rubocop-rails-omakase", require: false
# Standard Ruby style guide, linter, and formatter [https://github.com/standardrb/standard]
gem "standard", require: false
end
group :development do
@@ -86,4 +87,10 @@ group :test do
# Use system testing [https://guides.rubyonrails.org/testing.html#system-testing]
gem "capybara"
gem "selenium-webdriver"
# Code coverage analysis
gem "simplecov", require: false
# Pin minitest to < 6.0 until Rails 8.1 supports the new API
gem "minitest", "< 6.0"
end

View File

@@ -1,7 +1,7 @@
GEM
remote: https://rubygems.org/
specs:
action_text-trix (2.1.15)
action_text-trix (2.1.16)
railties
actioncable (8.1.1)
actionpack (= 8.1.1)
@@ -75,23 +75,23 @@ GEM
securerandom (>= 0.3)
tzinfo (~> 2.0, >= 2.0.5)
uri (>= 0.13.1)
addressable (2.8.7)
public_suffix (>= 2.0.2, < 7.0)
addressable (2.8.8)
public_suffix (>= 2.0.2, < 8.0)
android_key_attestation (0.3.0)
ast (2.4.3)
base64 (0.3.0)
bcrypt (3.1.20)
bcrypt_pbkdf (1.1.1)
bigdecimal (3.3.1)
bcrypt (3.1.21)
bcrypt_pbkdf (1.1.2)
bigdecimal (4.0.1)
bindata (2.5.1)
bindex (0.8.1)
bootsnap (1.18.6)
bootsnap (1.20.1)
msgpack (~> 1.2)
brakeman (7.1.0)
brakeman (7.1.2)
racc
builder (3.3.0)
bundler-audit (0.9.2)
bundler (>= 1.2.0, < 3)
bundler-audit (0.9.3)
bundler (>= 1.2.0)
thor (~> 1.0)
capybara (3.40.0)
addressable
@@ -106,31 +106,37 @@ GEM
childprocess (5.1.0)
logger (~> 1.5)
chunky_png (1.4.0)
concurrent-ruby (1.3.5)
connection_pool (2.5.4)
concurrent-ruby (1.3.6)
connection_pool (3.0.2)
cose (1.3.1)
cbor (~> 0.5.9)
openssl-signature_algorithm (~> 1.0)
crass (1.0.6)
date (3.5.0)
debug (1.11.0)
date (3.5.1)
debug (1.11.1)
irb (~> 1.10)
reline (>= 0.3.8)
dotenv (3.1.8)
docile (1.4.1)
dotenv (3.2.0)
drb (2.2.3)
ed25519 (1.4.0)
erb (5.1.3)
erb (6.0.1)
erubi (1.13.1)
ffi (1.17.2-aarch64-linux-gnu)
ffi (1.17.2-aarch64-linux-musl)
ffi (1.17.2-arm-linux-gnu)
ffi (1.17.2-arm-linux-musl)
ffi (1.17.2-arm64-darwin)
ffi (1.17.2-x86_64-linux-gnu)
ffi (1.17.2-x86_64-linux-musl)
et-orbi (1.4.0)
tzinfo
ffi (1.17.3-aarch64-linux-gnu)
ffi (1.17.3-aarch64-linux-musl)
ffi (1.17.3-arm-linux-gnu)
ffi (1.17.3-arm-linux-musl)
ffi (1.17.3-arm64-darwin)
ffi (1.17.3-x86_64-linux-gnu)
ffi (1.17.3-x86_64-linux-musl)
fugit (1.12.1)
et-orbi (~> 1.4)
raabro (~> 1.4)
globalid (1.3.0)
activesupport (>= 6.1)
i18n (1.14.7)
i18n (1.14.8)
concurrent-ruby (~> 1.0)
image_processing (1.14.0)
mini_magick (>= 4.9.5, < 6)
@@ -139,18 +145,18 @@ GEM
actionpack (>= 6.0.0)
activesupport (>= 6.0.0)
railties (>= 6.0.0)
io-console (0.8.1)
irb (1.15.3)
io-console (0.8.2)
irb (1.16.0)
pp (>= 0.6.0)
rdoc (>= 4.0.0)
reline (>= 0.4.2)
jbuilder (2.14.1)
actionview (>= 7.0.0)
activesupport (>= 7.0.0)
json (2.15.2)
json (2.18.0)
jwt (3.1.2)
base64
kamal (2.8.1)
kamal (2.10.1)
activesupport (>= 7.0)
base64 (~> 0.2)
bcrypt_pbkdf (~> 1.0)
@@ -170,7 +176,7 @@ GEM
launchy (>= 2.2, < 4)
lint_roller (1.1.0)
logger (1.7.0)
loofah (2.24.1)
loofah (2.25.0)
crass (~> 1.0.2)
nokogiri (>= 1.12.0)
mail (2.9.0)
@@ -184,9 +190,9 @@ GEM
mini_magick (5.3.1)
logger
mini_mime (1.1.5)
minitest (5.26.0)
minitest (5.27.0)
msgpack (1.8.0)
net-imap (0.5.12)
net-imap (0.6.2)
date
net-protocol
net-pop (0.1.2)
@@ -201,42 +207,43 @@ GEM
net-protocol
net-ssh (7.3.0)
nio4r (2.7.5)
nokogiri (1.18.10-aarch64-linux-gnu)
nokogiri (1.19.0-aarch64-linux-gnu)
racc (~> 1.4)
nokogiri (1.18.10-aarch64-linux-musl)
nokogiri (1.19.0-aarch64-linux-musl)
racc (~> 1.4)
nokogiri (1.18.10-arm-linux-gnu)
nokogiri (1.19.0-arm-linux-gnu)
racc (~> 1.4)
nokogiri (1.18.10-arm-linux-musl)
nokogiri (1.19.0-arm-linux-musl)
racc (~> 1.4)
nokogiri (1.18.10-arm64-darwin)
nokogiri (1.19.0-arm64-darwin)
racc (~> 1.4)
nokogiri (1.18.10-x86_64-linux-gnu)
nokogiri (1.19.0-x86_64-linux-gnu)
racc (~> 1.4)
nokogiri (1.18.10-x86_64-linux-musl)
nokogiri (1.19.0-x86_64-linux-musl)
racc (~> 1.4)
openssl (3.3.2)
openssl (4.0.0)
openssl-signature_algorithm (1.3.0)
openssl (> 2.0)
ostruct (0.6.3)
parallel (1.27.0)
parser (3.3.9.0)
parser (3.3.10.0)
ast (~> 2.4.1)
racc
pp (0.6.3)
prettyprint
prettyprint (0.2.0)
prism (1.6.0)
prism (1.7.0)
propshaft (1.3.1)
actionpack (>= 7.0.0)
activesupport (>= 7.0.0)
rack
psych (5.2.6)
psych (5.3.1)
date
stringio
public_suffix (6.0.2)
public_suffix (7.0.0)
puma (7.1.0)
nio4r (~> 2.0)
raabro (1.4.0)
racc (1.8.1)
rack (3.2.4)
rack-session (2.1.1)
@@ -244,7 +251,7 @@ GEM
rack (>= 3.0.0)
rack-test (2.2.0)
rack (>= 1.3)
rackup (2.2.1)
rackup (2.3.1)
rack (>= 3)
rails (8.1.1)
actioncable (= 8.1.1)
@@ -278,20 +285,20 @@ GEM
zeitwerk (~> 2.6)
rainbow (3.1.1)
rake (13.3.1)
rdoc (6.15.1)
rdoc (7.0.3)
erb
psych (>= 4.0.0)
tsort
regexp_parser (2.11.3)
reline (0.6.2)
reline (0.6.3)
io-console (~> 0.5)
rexml (3.4.4)
rotp (6.3.0)
rqrcode (3.1.0)
rqrcode (3.1.1)
chunky_png (~> 1.0)
rqrcode_core (~> 2.0)
rqrcode_core (2.0.0)
rubocop (1.81.6)
rqrcode_core (2.0.1)
rubocop (1.81.7)
json (~> 2.3)
language_server-protocol (~> 3.17.0.2)
lint_roller (~> 1.1.0)
@@ -302,98 +309,113 @@ GEM
rubocop-ast (>= 1.47.1, < 2.0)
ruby-progressbar (~> 1.7)
unicode-display_width (>= 2.4.0, < 4.0)
rubocop-ast (1.47.1)
rubocop-ast (1.49.0)
parser (>= 3.3.7.2)
prism (~> 1.4)
prism (~> 1.7)
rubocop-performance (1.26.1)
lint_roller (~> 1.1)
rubocop (>= 1.75.0, < 2.0)
rubocop-ast (>= 1.47.1, < 2.0)
rubocop-rails (2.33.4)
activesupport (>= 4.2.0)
lint_roller (~> 1.1)
rack (>= 1.1)
rubocop (>= 1.75.0, < 2.0)
rubocop-ast (>= 1.44.0, < 2.0)
rubocop-rails-omakase (1.1.0)
rubocop (>= 1.72)
rubocop-performance (>= 1.24)
rubocop-rails (>= 2.30)
ruby-progressbar (1.13.0)
ruby-vips (2.2.5)
ruby-vips (2.3.0)
ffi (~> 1.12)
logger
rubyzip (3.2.1)
rubyzip (3.2.2)
safety_net_attestation (0.5.0)
jwt (>= 2.0, < 4.0)
securerandom (0.4.1)
selenium-webdriver (4.38.0)
selenium-webdriver (4.39.0)
base64 (~> 0.2)
logger (~> 1.4)
rexml (~> 3.2, >= 3.2.5)
rubyzip (>= 1.2.2, < 4.0)
websocket (~> 1.0)
sentry-rails (5.28.0)
railties (>= 5.0)
sentry-ruby (~> 5.28.0)
sentry-ruby (5.28.0)
sentry-rails (6.2.0)
railties (>= 5.2.0)
sentry-ruby (~> 6.2.0)
sentry-ruby (6.2.0)
bigdecimal
concurrent-ruby (~> 1.0, >= 1.0.2)
simplecov (0.22.0)
docile (~> 1.1)
simplecov-html (~> 0.11)
simplecov_json_formatter (~> 0.1)
simplecov-html (0.13.2)
simplecov_json_formatter (0.1.4)
solid_cable (3.0.12)
actioncable (>= 7.2)
activejob (>= 7.2)
activerecord (>= 7.2)
railties (>= 7.2)
solid_cache (1.0.8)
solid_cache (1.0.10)
activejob (>= 7.2)
activerecord (>= 7.2)
railties (>= 7.2)
sqlite3 (2.7.4-aarch64-linux-gnu)
sqlite3 (2.7.4-aarch64-linux-musl)
sqlite3 (2.7.4-arm-linux-gnu)
sqlite3 (2.7.4-arm-linux-musl)
sqlite3 (2.7.4-arm64-darwin)
sqlite3 (2.7.4-x86_64-linux-gnu)
sqlite3 (2.7.4-x86_64-linux-musl)
sshkit (1.24.0)
solid_queue (1.2.4)
activejob (>= 7.1)
activerecord (>= 7.1)
concurrent-ruby (>= 1.3.1)
fugit (~> 1.11)
railties (>= 7.1)
thor (>= 1.3.1)
sqlite3 (2.9.0-aarch64-linux-gnu)
sqlite3 (2.9.0-aarch64-linux-musl)
sqlite3 (2.9.0-arm-linux-gnu)
sqlite3 (2.9.0-arm-linux-musl)
sqlite3 (2.9.0-arm64-darwin)
sqlite3 (2.9.0-x86_64-linux-gnu)
sqlite3 (2.9.0-x86_64-linux-musl)
sshkit (1.25.0)
base64
logger
net-scp (>= 1.1.2)
net-sftp (>= 2.1.2)
net-ssh (>= 2.8.0)
ostruct
standard (1.52.0)
language_server-protocol (~> 3.17.0.2)
lint_roller (~> 1.0)
rubocop (~> 1.81.7)
standard-custom (~> 1.0.0)
standard-performance (~> 1.8)
standard-custom (1.0.2)
lint_roller (~> 1.0)
rubocop (~> 1.50)
standard-performance (1.9.0)
lint_roller (~> 1.1)
rubocop-performance (~> 1.26.0)
stimulus-rails (1.3.4)
railties (>= 6.0.0)
stringio (3.1.7)
tailwindcss-rails (4.3.0)
stringio (3.2.0)
tailwindcss-rails (4.4.0)
railties (>= 7.0.0)
tailwindcss-ruby (~> 4.0)
tailwindcss-ruby (4.1.13)
tailwindcss-ruby (4.1.13-aarch64-linux-gnu)
tailwindcss-ruby (4.1.13-aarch64-linux-musl)
tailwindcss-ruby (4.1.13-arm64-darwin)
tailwindcss-ruby (4.1.13-x86_64-linux-gnu)
tailwindcss-ruby (4.1.13-x86_64-linux-musl)
tailwindcss-ruby (4.1.18)
tailwindcss-ruby (4.1.18-aarch64-linux-gnu)
tailwindcss-ruby (4.1.18-aarch64-linux-musl)
tailwindcss-ruby (4.1.18-arm64-darwin)
tailwindcss-ruby (4.1.18-x86_64-linux-gnu)
tailwindcss-ruby (4.1.18-x86_64-linux-musl)
thor (1.4.0)
thruster (0.1.16)
thruster (0.1.16-aarch64-linux)
thruster (0.1.16-arm64-darwin)
thruster (0.1.16-x86_64-linux)
timeout (0.4.4)
thruster (0.1.17)
thruster (0.1.17-aarch64-linux)
thruster (0.1.17-arm64-darwin)
thruster (0.1.17-x86_64-linux)
timeout (0.6.0)
tpm-key_attestation (0.14.1)
bindata (~> 2.4)
openssl (> 2.0)
openssl-signature_algorithm (~> 1.0)
tsort (0.2.0)
turbo-rails (2.0.17)
turbo-rails (2.0.20)
actionpack (>= 7.1.0)
railties (>= 7.1.0)
tzinfo (2.0.6)
concurrent-ruby (~> 1.0)
unicode-display_width (3.2.0)
unicode-emoji (~> 4.1)
unicode-emoji (4.1.0)
uri (1.1.0)
unicode-emoji (4.2.0)
uri (1.1.1)
useragent (0.16.11)
web-console (4.2.1)
actionview (>= 6.0.0)
@@ -415,7 +437,7 @@ GEM
websocket-extensions (0.1.5)
xpath (3.2.0)
nokogiri (~> 1.8)
zeitwerk (2.7.3)
zeitwerk (2.7.4)
PLATFORMS
aarch64-linux
@@ -441,19 +463,22 @@ DEPENDENCIES
jwt (~> 3.1)
kamal
letter_opener
minitest (< 6.0)
propshaft
public_suffix (~> 6.0)
public_suffix (~> 7.0)
puma (>= 5.0)
rails (~> 8.1.1)
rotp (~> 6.3)
rqrcode (~> 3.1)
rubocop-rails-omakase
selenium-webdriver
sentry-rails (~> 5.18)
sentry-ruby (~> 5.18)
sentry-rails (~> 6.2)
sentry-ruby (~> 6.2)
simplecov
solid_cable
solid_cache
solid_queue (~> 1.2)
sqlite3 (>= 2.1)
standard
stimulus-rails
tailwindcss-rails
thruster
@@ -463,4 +488,4 @@ DEPENDENCIES
webauthn (~> 3.0)
BUNDLED WITH
2.7.2
4.0.3

587
README.md
View File

@@ -1,32 +1,17 @@
# Clinch
> [!NOTE]
> This software is experiemental. If you'd like to try it out, find bugs, security flaws and improvements, please do.
> This software is experimental. If you'd like to try it out, find bugs, security flaws and improvements, please do.
**A lightweight, self-hosted identity & SSO / IpD portal**
Clinch gives you one place to manage users and lets any web app authenticate against it without maintaining its own user table.
I've completed all planned features:
* Create Admin user on first login
* TOTP ( QR Code ) 2FA, with backup codes ( encrypted at rest )
* Passkey generation and login, with detection of Passkey during login
* Forward Auth configured and working
* OIDC provider with auto discovery, refresh tokens, and token revocation
* Configurable token expiry per application (access, refresh, ID tokens)
* Invite users by email, assign to groups
* Self managed password reset by email
* Use Groups to assign Applications ( Family group can access Kavita, Developers can access Gitea )
* Configurable Group and User custom claims for OIDC token
* Display all Applications available to the user on their Dashboard
* Display all logged in sessions and OIDC logged in sessions
What remains now is ensure test coverage,
Clinch gives you one place to manage users and lets any web app authenticate against it without managing its own users.
## Why Clinch?
Do you host your own web apps? MeTube, Kavita, Audiobookshelf, Gitea? Rather than managing all those separate user accounts, set everyone up on Clinch and let it do the authentication and user management.
Do you host your own web apps? MeTube, Kavita, Audiobookshelf, Gitea, Grafana, Proxmox? Rather than managing all those separate user accounts, set everyone up on Clinch and let it do the authentication and user management.
Clinch runs as a single Docker container, using SQLite as the database, the job queue (Solid Queue) and the shared cache (Solid Cache). The webserver, Puma, runs the job queue in-process, avoiding the need for another container.
Clinch sits in a sweet spot between two excellent open-source identity solutions:
@@ -84,6 +69,9 @@ Clinch sits in a sweet spot between two excellent open-source identity solutions
### SSO Protocols
Apps that speak OIDC use the OIDC flow.
Apps that only need "who is it?", or you want available from the internet behind authentication (MeTube, Jellyfin) use ForwardAuth.
#### OpenID Connect (OIDC)
Standard OAuth2/OIDC provider with endpoints:
- `/.well-known/openid-configuration` - Discovery endpoint
@@ -94,19 +82,47 @@ Standard OAuth2/OIDC provider with endpoints:
Features:
- **Refresh tokens** - Long-lived tokens (30 days default) with automatic rotation and revocation
- **Token family tracking** - Advanced security detects token replay attacks and revokes compromised token families
- **Configurable token expiry** - Set access token (5min-24hr), refresh token (1-90 days), and ID token TTL per application
- **Token security** - BCrypt-hashed tokens, automatic cleanup of expired tokens
- **Token security** - All tokens HMAC-SHA256 hashed (suitable for 256-bit random data), automatic cleanup of expired tokens
- **Pairwise subject identifiers** - Each user gets a unique, stable `sub` claim per application for enhanced privacy
Client apps (Audiobookshelf, Kavita, Grafana, etc.) redirect to Clinch for login and receive ID tokens, access tokens, and refresh tokens.
**ID Token Claims** (JWT with RS256 signature):
| Claim | Description | Notes |
|-------|-------------|-------|
| Standard Claims | | |
| `iss` | Issuer (Clinch URL) | From `CLINCH_HOST` |
| `sub` | Subject (user identifier) | Pairwise SID - unique per app |
| `aud` | Audience | OAuth client_id |
| `exp` | Expiration timestamp | Configurable TTL |
| `iat` | Issued-at timestamp | Token creation time |
| `email` | User email | |
| `email_verified` | Email verification | Always `true` |
| `preferred_username` | Username/email | Fallback to email |
| `name` | Display name | User's name or email |
| `nonce` | Random value | From auth request (prevents replay) |
| **Security Claims** | | |
| `at_hash` | Access token hash | SHA-256 hash of access_token (OIDC Core §3.1.3.6) |
| `auth_time` | Authentication time | Unix timestamp of when user logged in (OIDC Core §2) |
| `acr` | Auth context class | `"1"` = password, `"2"` = 2FA/passkey (OIDC Core §2) |
| `azp` | Authorized party | OAuth client_id (OIDC Core §2) |
| Custom Claims | | |
| `groups` | User's groups | Array of group names |
| *custom* | Arbitrary key-values | From groups, users, or app-specific config |
**Authentication Context Class Reference (`acr`):**
- `"1"` - Something you know (password only)
- `"2"` - Two-factor or phishing-resistant (TOTP, backup codes, WebAuthn/passkey)
Client apps (Audiobookshelf, Kavita, Proxmox, Grafana, etc.) redirect to Clinch for login and receive ID tokens, access tokens, and refresh tokens.
#### Trusted-Header SSO (ForwardAuth)
Works with reverse proxies (Caddy, Traefik, Nginx):
1. Proxy sends every request to `/api/verify`
2. **200 OK** → Proxy injects headers (`Remote-User`, `Remote-Groups`, `Remote-Email`) and forwards to app
3. **401/403** → Proxy redirects to Clinch login; after login, user returns to original URL
Apps that speak OIDC use the OIDC flow; apps that only need "who is it?" headers use ForwardAuth.
2. Response handling:
- **200 OK** → Proxy injects headers (`Remote-User`, `Remote-Groups`, `Remote-Email`) and forwards to app
- **Any other status** → Proxy returns that response directly to client (typically 302 redirect to login page)
**Note:** ForwardAuth requires applications to run on the same domain as Clinch (e.g., `app.yourdomain.com` with Clinch at `auth.yourdomain.com`) for secure session cookie sharing. Take a look at Authentik if you need multi domain support.
@@ -114,7 +130,6 @@ Apps that speak OIDC use the OIDC flow; apps that only need "who is it?" headers
Send emails for:
- Invitation links (one-time token, 7-day expiry)
- Password reset links (one-time token, 1-hour expiry)
- 2FA backup codes
### Session Management
- **Device tracking** - See all active sessions with device names and IPs
@@ -214,9 +229,9 @@ Configure different claims for different applications on a per-user basis:
- Many-to-many with Groups (allowlist)
**OIDC Tokens**
- Authorization codes (10-minute expiry, one-time use, PKCE support)
- Access tokens (opaque, BCrypt-hashed, configurable expiry 5min-24hr, revocable)
- Refresh tokens (opaque, BCrypt-hashed, configurable expiry 1-90 days, single-use with rotation)
- Authorization codes (opaque, HMAC-SHA256 hashed, 10-minute expiry, one-time use, PKCE support)
- Access tokens (opaque, HMAC-SHA256 hashed, configurable expiry 5min-24hr, revocable)
- Refresh tokens (opaque, HMAC-SHA256 hashed, configurable expiry 1-90 days, single-use with rotation)
- ID tokens (JWT, signed with RS256, configurable expiry 5min-24hr)
---
@@ -244,6 +259,24 @@ Configure different claims for different applications on a per-user basis:
- Proxy redirects to Clinch login page
- After login, redirect back to original URL
#### Race Condition Handling
After successful login, you may notice an `fa_token` query parameter appended to redirect URLs (e.g., `https://app.example.com/dashboard?fa_token=...`). This solves a timing issue:
**The Problem:**
1. User signs in → session cookie is set
2. Browser gets redirected to protected resource
3. Browser may not have processed the `Set-Cookie` header yet
4. Reverse proxy checks `/api/verify` → no cookie yet → auth fails ❌
**The Solution:**
- A one-time token (`fa_token`) is added to the redirect URL as a query parameter
- `/api/verify` checks for this token first, before checking cookies
- Token is cached for 60 seconds and deleted immediately after use
- This gives the browser's cookie handling time to catch up
This is transparent to end users and requires no configuration.
---
## Setup & Installation
@@ -269,56 +302,207 @@ bin/rails db:migrate
bin/dev
```
### Docker Deployment
---
## Production Deployment
### Docker Compose (Recommended)
Create a `docker-compose.yml` file:
```yaml
services:
clinch:
image: ghcr.io/dkam/clinch:latest
ports:
- "127.0.0.1:3000:3000" # Bind to localhost only (reverse proxy on same host)
# Use "3000:3000" if reverse proxy is in Docker network or different host
environment:
# Rails Configuration
RAILS_ENV: production
SECRET_KEY_BASE: ${SECRET_KEY_BASE}
# Application Configuration
CLINCH_HOST: ${CLINCH_HOST}
CLINCH_FROM_EMAIL: ${CLINCH_FROM_EMAIL:-noreply@example.com}
# SMTP Configuration
SMTP_ADDRESS: ${SMTP_ADDRESS}
SMTP_PORT: ${SMTP_PORT}
SMTP_DOMAIN: ${SMTP_DOMAIN}
SMTP_USERNAME: ${SMTP_USERNAME}
SMTP_PASSWORD: ${SMTP_PASSWORD}
SMTP_AUTHENTICATION: ${SMTP_AUTHENTICATION:-plain}
SMTP_ENABLE_STARTTLS: ${SMTP_ENABLE_STARTTLS:-true}
# OIDC Configuration (optional - generates temporary key if not provided)
OIDC_PRIVATE_KEY: ${OIDC_PRIVATE_KEY}
# Optional Configuration
FORCE_SSL: ${FORCE_SSL:-false}
volumes:
- ./storage:/rails/storage
restart: unless-stopped
```
Create a `.env` file in the same directory:
**Generate required secrets first:**
```bash
# Build image
docker build -t clinch .
# Generate SECRET_KEY_BASE (required)
openssl rand -hex 64
# Run container
docker run -p 3000:3000 \
-v clinch-storage:/rails/storage \
-e SECRET_KEY_BASE=your-secret-key \
-e SMTP_ADDRESS=smtp.example.com \
-e SMTP_PORT=587 \
-e SMTP_USERNAME=your-username \
-e SMTP_PASSWORD=your-password \
clinch
# Generate OIDC private key (optional - auto-generated if not provided)
openssl genpkey -algorithm RSA -out private_key.pem -pkeyopt rsa_keygen_bits:2048
cat private_key.pem # Copy the output into OIDC_PRIVATE_KEY below
```
**Then create `.env`:**
```bash
# Rails Secret (REQUIRED)
SECRET_KEY_BASE=paste-output-from-openssl-rand-hex-64-here
# Application URLs (REQUIRED)
CLINCH_HOST=https://auth.yourdomain.com
CLINCH_FROM_EMAIL=noreply@yourdomain.com
# SMTP Settings (REQUIRED for invitations and password resets)
SMTP_ADDRESS=smtp.example.com
SMTP_PORT=587
SMTP_DOMAIN=yourdomain.com
SMTP_USERNAME=your-smtp-username
SMTP_PASSWORD=your-smtp-password
# OIDC Private Key (OPTIONAL - generates temporary key if not provided)
# For production, generate a persistent key and paste the ENTIRE contents here
OIDC_PRIVATE_KEY=
# Optional: Force SSL redirects (only if NOT behind a reverse proxy handling SSL)
FORCE_SSL=false
```
Start Clinch:
```bash
docker compose up -d
```
**First Run:**
1. Visit `http://localhost:3000` (or your configured domain)
2. Complete the first-run wizard to create your admin account
3. Configure applications and invite users
**Upgrading:**
```bash
# Pull latest image
docker compose pull
# Restart with new image (migrations run automatically)
docker compose up -d
```
**Logs:**
```bash
# View logs
docker compose logs -f clinch
# View last 100 lines
docker compose logs --tail=100 clinch
```
### Backup & Restore
Clinch stores all persistent data in the `storage/` directory (or `/rails/storage` in Docker):
- SQLite database (`production.sqlite3`)
- Uploaded files via ActiveStorage (application icons)
**Database Backup:**
Use SQLite's `VACUUM INTO` command for safe, atomic backups of a running database:
```bash
# Local development
sqlite3 storage/production.sqlite3 "VACUUM INTO 'backup.sqlite3';"
```
This creates an optimized copy of the database that's safe to make even while Clinch is running.
**Full Backup (Database + Uploads):**
For complete backups including uploaded files, backup the database and uploads separately:
```bash
# 1. Backup database (safe while running)
sqlite3 storage/production.sqlite3 "VACUUM INTO 'backup-$(date +%Y%m%d).sqlite3';"
# 2. Backup uploaded files (ActiveStorage files are immutable)
tar -czf uploads-backup-$(date +%Y%m%d).tar.gz storage/uploads/
# Docker Compose equivalent
docker compose exec clinch sqlite3 /rails/storage/production.sqlite3 "VACUUM INTO '/rails/storage/backup-$(date +%Y%m%d).sqlite3';"
docker compose exec clinch tar -czf /rails/storage/uploads-backup-$(date +%Y%m%d).tar.gz /rails/storage/uploads/
```
**Restore:**
```bash
# Stop Clinch first
# Then restore database
cp backup-YYYYMMDD.sqlite3 storage/production.sqlite3
# Restore uploads
tar -xzf uploads-backup-YYYYMMDD.tar.gz -C storage/
```
**Docker Volume Backup:**
**Option 1: While Running (Online Backup)**
a) **Mapped volumes** (recommended, e.g., `-v /host/path:/rails/storage`):
```bash
# Database backup (safe while running)
sqlite3 /host/path/production.sqlite3 "VACUUM INTO '/host/path/backup-$(date +%Y%m%d).sqlite3';"
# Then sync to off-server storage
rsync -av /host/path/backup-*.sqlite3 /host/path/uploads/ remote:/backups/clinch/
```
b) **Docker volumes** (e.g., using named volumes in compose):
```bash
# Database backup (safe while running)
docker compose exec clinch sqlite3 /rails/storage/production.sqlite3 "VACUUM INTO '/rails/storage/backup.sqlite3';"
# Copy out of container
docker compose cp clinch:/rails/storage/backup.sqlite3 ./backup-$(date +%Y%m%d).sqlite3
```
**Option 2: While Stopped (Offline Backup)**
If Docker is stopped, you can copy the entire storage:
```bash
docker compose down
# For mapped volumes
tar -czf clinch-backup-$(date +%Y%m%d).tar.gz /host/path/
# For docker volumes
docker run --rm -v clinch_storage:/data -v $(pwd):/backup ubuntu \
tar czf /backup/clinch-backup-$(date +%Y%m%d).tar.gz /data
docker compose up -d
```
**Important:** Do not use tar/snapshots on a running database - use `VACUUM INTO` instead or stop the container first.
---
## Configuration
### Environment Variables
Create a `.env` file (see `.env.example`):
```bash
# Rails
SECRET_KEY_BASE=generate-with-bin-rails-secret
RAILS_ENV=production
# Database
# SQLite database stored in storage/ directory (Docker volume mount point)
# SMTP (for sending emails)
SMTP_ADDRESS=smtp.example.com
SMTP_PORT=587
SMTP_DOMAIN=example.com
SMTP_USERNAME=your-username
SMTP_PASSWORD=your-password
SMTP_AUTHENTICATION=plain
SMTP_ENABLE_STARTTLS=true
# Application
CLINCH_HOST=https://auth.example.com
CLINCH_FROM_EMAIL=noreply@example.com
# OIDC (optional - generates temporary key in development)
# Generate with: openssl genpkey -algorithm RSA -out private_key.pem -pkeyopt rsa_keygen_bits:2048
OIDC_PRIVATE_KEY=<contents-of-private-key.pem>
```
All configuration is handled via environment variables (see the `.env` file in the Docker Compose section above).
### First Run
1. Visit Clinch at `http://localhost:3000` (or your configured domain)
@@ -331,24 +515,255 @@ OIDC_PRIVATE_KEY=<contents-of-private-key.pem>
---
## Roadmap
## Rails Console
### In Progress
- OIDC provider implementation
- ForwardAuth endpoint
- Admin UI for user/group/app management
- First-run wizard
One advantage of being a Rails application is direct access to the Rails console for administrative tasks. This is particularly useful for debugging, emergency access, or bulk operations.
### Planned Features
- **Audit logging** - Track all authentication events
- **WebAuthn/Passkeys** - Hardware key support
### Starting the Console
#### Maybe
- **SAML support** - SAML 2.0 identity provider
- **Policy engine** - Rule-based access control
- Example: `IF user.email =~ "*@gmail.com" AND app.slug == "kavita" THEN DENY`
- Stored as JSON, evaluated after auth but before consent
- **LDAP sync** - Import users from LDAP/Active Directory
```bash
# Docker / Docker Compose
docker exec -it clinch bin/rails console
# or
docker compose exec -it clinch bin/rails console
# Local development
bin/rails console
```
### Finding Users
```ruby
# Find by email
user = User.find_by(email_address: 'alice@example.com')
# Find by username
user = User.find_by(username: 'alice')
# List all users
User.all.pluck(:id, :email_address, :status)
# Find admins
User.admins.pluck(:email_address)
# Find users in a specific status
User.active.count
User.disabled.pluck(:email_address)
User.pending_invitation.pluck(:email_address)
```
### Creating Users
```ruby
# Create a regular user
User.create!(
email_address: 'newuser@example.com',
password: 'secure-password-here',
status: :active
)
# Create an admin user
User.create!(
email_address: 'admin@example.com',
password: 'secure-password-here',
status: :active,
admin: true
)
```
### Managing Passwords
```ruby
user = User.find_by(email_address: 'alice@example.com')
user.password = 'new-secure-password'
user.save!
```
### Two-Factor Authentication (TOTP)
```ruby
user = User.find_by(email_address: 'alice@example.com')
# Check if TOTP is enabled
user.totp_enabled?
# Get current TOTP code (useful for testing/debugging)
puts user.console_totp
# Enable TOTP (generates secret and backup codes)
backup_codes = user.enable_totp!
puts backup_codes # Display backup codes to give to user
# Disable TOTP
user.disable_totp!
# Force user to set up TOTP on next login
user.update!(totp_required: true)
```
### Managing User Status
```ruby
user = User.find_by(email_address: 'alice@example.com')
# Disable a user (prevents login)
user.disabled!
# Re-enable a user
user.active!
# Check current status
user.status # => "active", "disabled", or "pending_invitation"
# Grant admin privileges
user.update!(admin: true)
# Revoke admin privileges
user.update!(admin: false)
```
### Managing Groups
```ruby
user = User.find_by(email_address: 'alice@example.com')
# View user's groups
user.groups.pluck(:name)
# Add user to a group
family = Group.find_by(name: 'family')
user.groups << family
# Remove user from a group
user.groups.delete(family)
# Create a new group
Group.create!(name: 'developers', description: 'Development team')
```
### Managing Sessions
```ruby
user = User.find_by(email_address: 'alice@example.com')
# View active sessions
user.sessions.pluck(:id, :device_name, :client_ip, :created_at)
# Revoke all sessions (force logout everywhere)
user.sessions.destroy_all
# Revoke a specific session
user.sessions.find(123).destroy
```
### Managing Applications
```ruby
# List all OIDC applications
Application.oidc.pluck(:name, :client_id)
# Find an application
app = Application.find_by(slug: 'kavita')
# Regenerate client secret
new_secret = app.generate_new_client_secret!
puts new_secret # Display once - not stored in plain text
# Check which users can access an app
app.allowed_groups.flat_map(&:users).uniq.pluck(:email_address)
# Revoke all tokens for an application
app.oidc_access_tokens.destroy_all
app.oidc_refresh_tokens.destroy_all
```
### Revoking OIDC Consents
```ruby
user = User.find_by(email_address: 'alice@example.com')
app = Application.find_by(slug: 'kavita')
# Revoke consent for a specific app
user.revoke_consent!(app)
# Revoke all OIDC consents
user.revoke_all_consents!
```
---
## Testing & Security
### Running Tests
Clinch has comprehensive test coverage with 341 tests covering integration, models, controllers, services, and system tests.
```bash
# Run all tests
bin/rails test
# Run specific test types
bin/rails test:integration
bin/rails test:models
bin/rails test:controllers
bin/rails test:system
# Run with code coverage report
COVERAGE=1 bin/rails test
# View coverage report at coverage/index.html
```
### Security Scanning
Clinch uses multiple automated security tools to ensure code quality and security:
```bash
# Run all security checks
bin/rake security
# Individual security scans
bin/brakeman --no-pager # Static security analysis
bin/bundler-audit check --update # Dependency vulnerability scan
bin/importmap audit # JavaScript dependency scan
```
**Container Image Scanning:**
```bash
# Install Trivy
brew install trivy # macOS
# or use Docker: alias trivy='docker run --rm -v /var/run/docker.sock:/var/run/docker.sock aquasec/trivy'
# Build and scan image (CRITICAL and HIGH severity only, like CI)
docker build -t clinch:local .
trivy image --severity CRITICAL,HIGH --scanners vuln clinch:local
# Scan only for fixable vulnerabilities
trivy image --severity CRITICAL,HIGH --scanners vuln --ignore-unfixed clinch:local
```
**CI/CD Integration:**
All security scans run automatically on every pull request and push to main via GitHub Actions.
**Security Tools:**
- **Brakeman** - Static analysis for Rails security vulnerabilities
- **bundler-audit** - Checks gems for known CVEs
- **Trivy** - Container image vulnerability scanning (OS/system packages)
- **Dependabot** - Automated dependency updates
- **GitHub Secret Scanning** - Detects leaked credentials with push protection
- **SimpleCov** - Code coverage tracking
- **RuboCop** - Code style and quality enforcement
**Current Status:**
- ✅ All security scans passing
- ✅ 341 tests, 1349 assertions, 0 failures
- ✅ No known dependency vulnerabilities
- ✅ Phases 1-4 security hardening complete (18+ vulnerabilities fixed)
- 🟡 3 outstanding security issues (all MEDIUM/LOW priority)
**Security Documentation:**
- [docs/security-todo.md](docs/security-todo.md) - Detailed vulnerability tracking and remediation history
- [docs/beta-checklist.md](docs/beta-checklist.md) - Beta release readiness criteria
---

View File

@@ -1 +0,0 @@
2025.03

View File

@@ -7,8 +7,9 @@ module ApplicationCable
end
private
def set_current_user
if session = Session.find_by(id: cookies.signed[:session_id])
if (session = Session.find_by(id: cookies.signed[:session_id]))
self.current_user = session.user
end
end

View File

@@ -16,16 +16,82 @@ class ActiveSessionsController < ApplicationController
return
end
# Send backchannel logout notification before revoking consent
if application.supports_backchannel_logout?
BackchannelLogoutJob.perform_later(
user_id: @user.id,
application_id: application.id,
consent_sid: consent.sid
)
Rails.logger.info "ActiveSessionsController: Enqueued backchannel logout for #{application.name}"
end
# Revoke all tokens for this user-application pair
now = Time.current
revoked_access_tokens = OidcAccessToken.where(application: application, user: @user, revoked_at: nil)
.update_all(revoked_at: now)
revoked_refresh_tokens = OidcRefreshToken.where(application: application, user: @user, revoked_at: nil)
.update_all(revoked_at: now)
Rails.logger.info "ActiveSessionsController: Revoked #{revoked_access_tokens} access tokens and #{revoked_refresh_tokens} refresh tokens for #{application.name}"
# Revoke the consent
consent.destroy
redirect_to active_sessions_path, notice: "Successfully revoked access to #{application.name}."
end
def logout_from_app
@user = Current.session.user
application = Application.find(params[:application_id])
# Check if user has consent for this application
consent = @user.oidc_user_consents.find_by(application: application)
unless consent
redirect_to root_path, alert: "No active session found for this application."
return
end
# Send backchannel logout notification
if application.supports_backchannel_logout?
BackchannelLogoutJob.perform_later(
user_id: @user.id,
application_id: application.id,
consent_sid: consent.sid
)
Rails.logger.info "ActiveSessionsController: Enqueued backchannel logout for #{application.name}"
end
# Revoke all tokens for this user-application pair
now = Time.current
revoked_access_tokens = OidcAccessToken.where(application: application, user: @user, revoked_at: nil)
.update_all(revoked_at: now)
revoked_refresh_tokens = OidcRefreshToken.where(application: application, user: @user, revoked_at: nil)
.update_all(revoked_at: now)
Rails.logger.info "ActiveSessionsController: Logged out from #{application.name} - revoked #{revoked_access_tokens} access tokens and #{revoked_refresh_tokens} refresh tokens"
# Keep the consent intact - this is the key difference from revoke_consent
redirect_to root_path, notice: "Successfully logged out of #{application.name}."
end
def revoke_all_consents
@user = Current.session.user
count = @user.oidc_user_consents.count
consents = @user.oidc_user_consents.includes(:application)
count = consents.count
if count > 0
# Send backchannel logout notifications before revoking consents
consents.each do |consent|
next unless consent.application.supports_backchannel_logout?
BackchannelLogoutJob.perform_later(
user_id: @user.id,
application_id: consent.application.id,
consent_sid: consent.sid
)
end
Rails.logger.info "ActiveSessionsController: Enqueued #{count} backchannel logout notifications"
@user.oidc_user_consents.destroy_all
redirect_to active_sessions_path, notice: "Successfully revoked access to #{count} applications."
else

View File

@@ -26,18 +26,17 @@ module Admin
@application.allowed_groups = Group.where(id: group_ids)
end
# Get the plain text client secret to show one time
# Get the plain text client secret to show one time (confidential clients only)
client_secret = nil
if @application.oidc?
if @application.oidc? && @application.confidential_client?
client_secret = @application.generate_new_client_secret!
end
if @application.oidc? && client_secret
flash[:notice] = "Application created successfully."
if @application.oidc?
flash[:client_id] = @application.client_id
flash[:client_secret] = client_secret
else
flash[:notice] = "Application created successfully."
flash[:client_secret] = client_secret if client_secret
flash[:public_client] = true if @application.public_client?
end
redirect_to admin_application_path(@application)
@@ -74,15 +73,20 @@ module Admin
def regenerate_credentials
if @application.oidc?
# Generate new client ID and secret
# Generate new client ID (always)
new_client_id = SecureRandom.urlsafe_base64(32)
client_secret = @application.generate_new_client_secret!
@application.update!(client_id: new_client_id)
flash[:notice] = "Credentials regenerated successfully."
flash[:client_id] = @application.client_id
# Generate new client secret only for confidential clients
if @application.confidential_client?
client_secret = @application.generate_new_client_secret!
flash[:client_secret] = client_secret
else
flash[:public_client] = true
end
redirect_to admin_application_path(@application)
else
@@ -97,14 +101,24 @@ module Admin
end
def application_params
params.require(:application).permit(
permitted = params.require(:application).permit(
:name, :slug, :app_type, :active, :redirect_uris, :description, :metadata,
:domain_pattern, :landing_url, :access_token_ttl, :refresh_token_ttl, :id_token_ttl,
headers_config: {}
).tap do |whitelisted|
# Remove client_secret from params if present (shouldn't be updated via form)
whitelisted.delete(:client_secret)
:icon, :backchannel_logout_uri, :is_public_client, :require_pkce
)
# Handle headers_config - it comes as a JSON string from the text area
if params[:application][:headers_config].present?
begin
permitted[:headers_config] = JSON.parse(params[:application][:headers_config])
rescue JSON::ParserError
permitted[:headers_config] = {}
end
end
# Remove client_secret from params if present (shouldn't be updated via form)
permitted.delete(:client_secret)
permitted
end
end
end

View File

@@ -8,7 +8,7 @@ module Api
def violation_report
# Parse CSP violation report
report_data = JSON.parse(request.body.read)
csp_report = report_data['csp-report']
csp_report = report_data["csp-report"]
# Validate that we have a proper CSP report
unless csp_report.is_a?(Hash) && csp_report.present?
@@ -19,28 +19,28 @@ module Api
# Log the violation for security monitoring
Rails.logger.warn "CSP Violation Report:"
Rails.logger.warn " Blocked URI: #{csp_report['blocked-uri']}"
Rails.logger.warn " Document URI: #{csp_report['document-uri']}"
Rails.logger.warn " Referrer: #{csp_report['referrer']}"
Rails.logger.warn " Violated Directive: #{csp_report['violated-directive']}"
Rails.logger.warn " Original Policy: #{csp_report['original-policy']}"
Rails.logger.warn " Blocked URI: #{csp_report["blocked-uri"]}"
Rails.logger.warn " Document URI: #{csp_report["document-uri"]}"
Rails.logger.warn " Referrer: #{csp_report["referrer"]}"
Rails.logger.warn " Violated Directive: #{csp_report["violated-directive"]}"
Rails.logger.warn " Original Policy: #{csp_report["original-policy"]}"
Rails.logger.warn " User Agent: #{request.user_agent}"
Rails.logger.warn " IP Address: #{request.remote_ip}"
# Emit structured event for CSP violation
# This allows multiple subscribers to process the event (Sentry, local logging, etc.)
Rails.event.notify("csp.violation", {
blocked_uri: csp_report['blocked-uri'],
document_uri: csp_report['document-uri'],
referrer: csp_report['referrer'],
violated_directive: csp_report['violated-directive'],
original_policy: csp_report['original-policy'],
disposition: csp_report['disposition'],
effective_directive: csp_report['effective-directive'],
source_file: csp_report['source-file'],
line_number: csp_report['line-number'],
column_number: csp_report['column-number'],
status_code: csp_report['status-code'],
blocked_uri: csp_report["blocked-uri"],
document_uri: csp_report["document-uri"],
referrer: csp_report["referrer"],
violated_directive: csp_report["violated-directive"],
original_policy: csp_report["original-policy"],
disposition: csp_report["disposition"],
effective_directive: csp_report["effective-directive"],
source_file: csp_report["source-file"],
line_number: csp_report["line-number"],
column_number: csp_report["column-number"],
status_code: csp_report["status-code"],
user_agent: request.user_agent,
ip_address: request.remote_ip,
current_user_id: Current.user&.id,

View File

@@ -3,7 +3,7 @@ module Api
# ForwardAuth endpoints need session storage for return URL
allow_unauthenticated_access
skip_before_action :verify_authenticity_token
rate_limit to: 100, within: 1.minute, only: :verify, with: -> { head :too_many_requests }
# No rate limiting on forward_auth endpoint - proxy middleware hits this frequently
# GET /api/verify
# This endpoint is called by reverse proxies (Traefik, Caddy, nginx)
@@ -49,14 +49,20 @@ module Api
forwarded_host = request.headers["X-Forwarded-Host"] || request.headers["Host"]
if forwarded_host.present?
# Load active forward auth applications with their associations for better performance
# Load all forward auth applications (including inactive ones) for security checks
# Preload groups to avoid N+1 queries in user_allowed? checks
apps = Application.forward_auth.includes(:allowed_groups).active
apps = Application.forward_auth.includes(:allowed_groups)
# Find matching forward auth application for this domain
app = apps.find { |a| a.matches_domain?(forwarded_host) }
if app
# Check if application is active
unless app.active?
Rails.logger.info "ForwardAuth: Access denied to #{forwarded_host} - application is inactive"
return render_forbidden("No authentication rule configured for this domain")
end
# Check if user is allowed by this application
unless app.user_allowed?(user)
Rails.logger.info "ForwardAuth: User #{user.email_address} denied access to #{forwarded_host} by app #{app.domain_pattern}"
@@ -65,8 +71,9 @@ module Api
Rails.logger.info "ForwardAuth: User #{user.email_address} granted access to #{forwarded_host} by app #{app.domain_pattern} (policy: #{app.policy_for_user(user)})"
else
# No application found - allow access with default headers (original behavior)
Rails.logger.info "ForwardAuth: No application found for domain: #{forwarded_host}, allowing with default headers"
# No application found - DENY by default (fail-closed security)
Rails.logger.info "ForwardAuth: Access denied to #{forwarded_host} - no authentication rule configured"
return render_forbidden("No authentication rule configured for this domain")
end
else
Rails.logger.info "ForwardAuth: User #{user.email_address} authenticated (no domain specified)"
@@ -74,7 +81,10 @@ module Api
# User is authenticated and authorized
# Return 200 with user information headers using app-specific configuration
headers = app ? app.headers_for_user(user) : Application::DEFAULT_HEADERS.map { |key, header_name|
headers = if app
app.headers_for_user(user)
else
Application::DEFAULT_HEADERS.map { |key, header_name|
case key
when :user, :email, :name
[header_name, user.email_address]
@@ -84,12 +94,13 @@ module Api
[header_name, user.admin? ? "true" : "false"]
end
}.compact.to_h
end
headers.each { |key, value| response.headers[key] = value }
# Log what headers we're sending (helpful for debugging)
if headers.any?
Rails.logger.debug "ForwardAuth: Headers sent: #{headers.keys.join(', ')}"
Rails.logger.debug "ForwardAuth: Headers sent: #{headers.keys.join(", ")}"
else
Rails.logger.debug "ForwardAuth: No headers sent (access only)"
end
@@ -122,8 +133,7 @@ module Api
def extract_session_id
# Extract session ID from cookie
# Rails uses signed cookies by default
session_id = cookies.signed[:session_id]
session_id
cookies.signed[:session_id]
end
def extract_app_from_headers
@@ -135,6 +145,9 @@ module Api
def render_unauthorized(reason = nil)
Rails.logger.info "ForwardAuth: Unauthorized - #{reason}"
# Set auth reason header for debugging (like Authelia)
response.headers["X-Auth-Reason"] = reason if reason.present?
# Get the redirect URL from query params or construct default
redirect_url = validate_redirect_url(params[:rd])
base_url = determine_base_url(redirect_url)
@@ -145,7 +158,7 @@ module Api
original_uri = request.headers["X-Forwarded-Uri"] || request.headers["X-Forwarded-Path"] || "/"
# Debug logging to see what headers we're getting
Rails.logger.info "ForwardAuth Headers: Host=#{request.headers['Host']}, X-Forwarded-Host=#{original_host}, X-Forwarded-Uri=#{request.headers['X-Forwarded-Uri']}, X-Forwarded-Path=#{request.headers['X-Forwarded-Path']}"
Rails.logger.info "ForwardAuth Headers: Host=#{request.headers["Host"]}, X-Forwarded-Host=#{original_host}, X-Forwarded-Uri=#{request.headers["X-Forwarded-Uri"]}, X-Forwarded-Path=#{request.headers["X-Forwarded-Path"]}"
original_url = if original_host
# Use the forwarded host and URI (original behavior)
@@ -176,6 +189,9 @@ module Api
def render_forbidden(reason = nil)
Rails.logger.info "ForwardAuth: Forbidden - #{reason}"
# Set auth reason header for debugging (like Authelia)
response.headers["X-Auth-Reason"] = reason if reason.present?
# Return 403 Forbidden
head :forbidden
end
@@ -190,7 +206,7 @@ module Api
return nil unless uri.is_a?(URI::HTTP) || uri.is_a?(URI::HTTPS)
# Only allow HTTPS in production
return nil unless Rails.env.development? || uri.scheme == 'https'
return nil unless Rails.env.development? || uri.scheme == "https"
redirect_domain = uri.host.downcase
return nil unless redirect_domain.present?
@@ -201,7 +217,6 @@ module Api
end
matching_app ? url : nil
rescue URI::InvalidURIError
nil
end
@@ -220,13 +235,13 @@ module Api
return redirect_url if redirect_url.present?
# Try CLINCH_HOST environment variable first
if ENV['CLINCH_HOST'].present?
host = ENV['CLINCH_HOST']
if ENV["CLINCH_HOST"].present?
host = ENV["CLINCH_HOST"]
# Ensure URL has https:// protocol
host.match?(/^https?:\/\//) ? host : "https://#{host}"
else
# Fallback to the request host
request_host = request.host || request.headers['X-Forwarded-Host']
request_host = request.host || request.headers["X-Forwarded-Host"]
if request_host.present?
Rails.logger.warn "ForwardAuth: CLINCH_HOST not set, using request host: #{request_host}"
"https://#{request_host}"

View File

@@ -1,5 +1,6 @@
class ApplicationController < ActionController::Base
include Authentication
# Only allow modern browsers supporting webp images, web push, badges, import maps, CSS nesting, and CSS :has.
allow_browser versions: :modern

View File

@@ -1,6 +1,6 @@
require 'uri'
require 'public_suffix'
require 'ipaddr'
require "uri"
require "public_suffix"
require "ipaddr"
module Authentication
extend ActiveSupport::Concern
@@ -17,6 +17,7 @@ module Authentication
end
private
def authenticated?
resume_session
end
@@ -39,14 +40,13 @@ module Authentication
end
def after_authentication_url
return_url = session[:return_to_after_authenticating]
final_url = session.delete(:return_to_after_authenticating) || root_url
final_url
session[:return_to_after_authenticating]
session.delete(:return_to_after_authenticating) || root_url
end
def start_new_session_for(user)
def start_new_session_for(user, acr: "1")
user.update!(last_sign_in_at: Time.current)
user.sessions.create!(user_agent: request.user_agent, ip_address: request.remote_ip).tap do |session|
user.sessions.create!(user_agent: request.user_agent, ip_address: request.remote_ip, acr: acr).tap do |session|
Current.session = session
# Extract root domain for cross-subdomain cookies (required for forward auth)
@@ -101,10 +101,14 @@ module Authentication
return nil if host.blank? || host.match?(/^(localhost|127\.0\.0\.1|::1)$/)
# Strip port number for domain parsing
host_without_port = host.split(':').first
host_without_port = host.split(":").first
# Check if it's an IP address (IPv4 or IPv6) - if so, don't set domain cookie
return nil if IPAddr.new(host_without_port) rescue false
begin
return nil if IPAddr.new(host_without_port)
rescue
false
end
# Use Public Suffix List for accurate domain parsing
domain = PublicSuffix.parse(host_without_port)
@@ -138,7 +142,7 @@ module Authentication
unless uri.path&.start_with?("/oauth/")
# Add token as query parameter
query_params = URI.decode_www_form(uri.query || "").to_h
query_params['fa_token'] = token
query_params["fa_token"] = token
uri.query = URI.encode_www_form(query_params)
# Update the session with the tokenized URL

View File

@@ -1,7 +1,9 @@
class InvitationsController < ApplicationController
include Authentication
allow_unauthenticated_access
before_action :set_user_by_invitation_token, only: %i[ show update ]
before_action :set_user_by_invitation_token, only: %i[show update]
rate_limit to: 10, within: 10.minutes, only: :update, with: -> { redirect_to signin_path, alert: "Too many attempts. Try again later." }
def show
# Show the password setup form
@@ -35,16 +37,16 @@ class InvitationsController < ApplicationController
# Check if user is still pending invitation
if @user.nil?
redirect_to signin_path, alert: "Invitation link is invalid or has expired."
return false
false
elsif @user.pending_invitation?
# User is valid and pending - proceed
return true
true
else
redirect_to signin_path, alert: "This invitation has already been used or is no longer valid."
return false
false
end
rescue ActiveSupport::MessageVerifier::InvalidSignature
redirect_to signin_path, alert: "Invitation link is invalid or has expired."
return false
false
end
end

View File

@@ -3,6 +3,14 @@ class OidcController < ApplicationController
allow_unauthenticated_access only: [:discovery, :jwks, :token, :revoke, :userinfo, :logout]
skip_before_action :verify_authenticity_token, only: [:token, :revoke, :logout]
# Rate limiting to prevent brute force and abuse
rate_limit to: 60, within: 1.minute, only: [:token, :revoke], with: -> {
render json: {error: "too_many_requests", error_description: "Rate limit exceeded. Try again later."}, status: :too_many_requests
}
rate_limit to: 30, within: 1.minute, only: [:authorize, :consent], with: -> {
render plain: "Too many authorization attempts. Try again later.", status: :too_many_requests
}
# GET /.well-known/openid-configuration
def discovery
base_url = OidcJwtService.issuer_url
@@ -18,12 +26,14 @@ class OidcController < ApplicationController
response_types_supported: ["code"],
response_modes_supported: ["query"],
grant_types_supported: ["authorization_code", "refresh_token"],
subject_types_supported: ["public"],
subject_types_supported: ["pairwise"],
id_token_signing_alg_values_supported: ["RS256"],
scopes_supported: ["openid", "profile", "email", "groups"],
scopes_supported: ["openid", "profile", "email", "groups", "offline_access"],
token_endpoint_auth_methods_supported: ["client_secret_post", "client_secret_basic"],
claims_supported: ["sub", "email", "email_verified", "name", "preferred_username", "groups", "admin"],
code_challenge_methods_supported: ["plain", "S256"]
claims_supported: ["sub", "email", "email_verified", "name", "preferred_username", "groups", "admin", "auth_time", "acr", "azp", "at_hash"],
code_challenge_methods_supported: ["plain", "S256"],
backchannel_logout_supported: true,
backchannel_logout_session_supported: true
}
render json: config
@@ -53,7 +63,7 @@ class OidcController < ApplicationController
error_details << "redirect_uri is required" unless redirect_uri.present?
error_details << "response_type must be 'code'" unless response_type == "code"
render plain: "Invalid request: #{error_details.join(', ')}", status: :bad_request
render plain: "Invalid request: #{error_details.join(", ")}", status: :bad_request
return
end
@@ -80,7 +90,7 @@ class OidcController < ApplicationController
Rails.logger.error "OAuth: Available OIDC applications: #{all_oidc_apps.pluck(:id, :client_id, :name)}"
error_msg = if Rails.env.development?
"Invalid request: Application not found for client_id '#{client_id}'. Available OIDC applications: #{all_oidc_apps.pluck(:name, :client_id).map { |name, id| "#{name} (#{id})" }.join(', ')}"
"Invalid request: Application not found for client_id '#{client_id}'. Available OIDC applications: #{all_oidc_apps.pluck(:name, :client_id).map { |name, id| "#{name} (#{id})" }.join(", ")}"
else
"Invalid request: Application not found"
end
@@ -89,13 +99,13 @@ class OidcController < ApplicationController
return
end
# Validate redirect URI
# Validate redirect URI first (required before we can safely redirect with errors)
unless @application.parsed_redirect_uris.include?(redirect_uri)
Rails.logger.error "OAuth: Invalid request - redirect URI mismatch. Expected: #{@application.parsed_redirect_uris}, Got: #{redirect_uri}"
# For development, show detailed error
error_msg = if Rails.env.development?
"Invalid request: Redirect URI mismatch. Application is configured for: #{@application.parsed_redirect_uris.join(', ')}, but received: #{redirect_uri}"
"Invalid request: Redirect URI mismatch. Application is configured for: #{@application.parsed_redirect_uris.join(", ")}, but received: #{redirect_uri}"
else
"Invalid request: Redirect URI not registered for this application"
end
@@ -104,6 +114,15 @@ class OidcController < ApplicationController
return
end
# Check if application is active (now we can safely redirect with error)
unless @application.active?
Rails.logger.error "OAuth: Application is not active: #{@application.name}"
error_uri = "#{redirect_uri}?error=unauthorized_client&error_description=Application+is+not+active"
error_uri += "&state=#{CGI.escape(state)}" if state.present?
redirect_to error_uri, allow_other_host: true
return
end
# Check if user is authenticated
unless authenticated?
# Store OAuth parameters in session and redirect to sign in
@@ -135,22 +154,22 @@ class OidcController < ApplicationController
existing_consent = user.has_oidc_consent?(@application, requested_scopes)
if existing_consent
# User has already consented, generate authorization code directly
code = SecureRandom.urlsafe_base64(32)
auth_code = OidcAuthorizationCode.create!(
application: @application,
user: user,
code: code,
redirect_uri: redirect_uri,
scope: scope,
nonce: nonce,
code_challenge: code_challenge,
code_challenge_method: code_challenge_method,
auth_time: Current.session.created_at.to_i,
acr: Current.session.acr,
expires_at: 10.minutes.from_now
)
# Redirect back to client with authorization code
redirect_uri = "#{redirect_uri}?code=#{code}"
redirect_uri += "&state=#{state}" if state.present?
# Redirect back to client with authorization code (plaintext)
redirect_uri = "#{redirect_uri}?code=#{auth_code.plaintext_code}"
redirect_uri += "&state=#{CGI.escape(state)}" if state.present?
redirect_to redirect_uri, allow_other_host: true
return
end
@@ -204,49 +223,55 @@ class OidcController < ApplicationController
# User denied consent
if params[:deny].present?
session.delete(:oauth_params)
error_uri = "#{oauth_params['redirect_uri']}?error=access_denied"
error_uri += "&state=#{oauth_params['state']}" if oauth_params['state']
error_uri = "#{oauth_params["redirect_uri"]}?error=access_denied"
error_uri += "&state=#{CGI.escape(oauth_params["state"])}" if oauth_params["state"]
redirect_to error_uri, allow_other_host: true
return
end
# Find the application
client_id = oauth_params['client_id']
client_id = oauth_params["client_id"]
application = Application.find_by(client_id: client_id, app_type: "oidc")
# Check if application is active (redirect with OAuth error)
unless application&.active?
Rails.logger.error "OAuth: Application is not active: #{application&.name || client_id}"
session.delete(:oauth_params)
error_uri = "#{oauth_params["redirect_uri"]}?error=unauthorized_client&error_description=Application+is+not+active"
error_uri += "&state=#{CGI.escape(oauth_params["state"])}" if oauth_params["state"].present?
redirect_to error_uri, allow_other_host: true
return
end
user = Current.session.user
# Record user consent
requested_scopes = oauth_params['scope'].split(' ')
OidcUserConsent.upsert(
{
user_id: user.id,
application_id: application.id,
scopes_granted: requested_scopes.join(' '),
granted_at: Time.current
},
unique_by: [:user_id, :application_id]
)
requested_scopes = oauth_params["scope"].split(" ")
consent = OidcUserConsent.find_or_initialize_by(user: user, application: application)
consent.scopes_granted = requested_scopes.join(" ")
consent.granted_at = Time.current
consent.save!
# Generate authorization code
code = SecureRandom.urlsafe_base64(32)
auth_code = OidcAuthorizationCode.create!(
application: application,
user: user,
code: code,
redirect_uri: oauth_params['redirect_uri'],
scope: oauth_params['scope'],
nonce: oauth_params['nonce'],
code_challenge: oauth_params['code_challenge'],
code_challenge_method: oauth_params['code_challenge_method'],
redirect_uri: oauth_params["redirect_uri"],
scope: oauth_params["scope"],
nonce: oauth_params["nonce"],
code_challenge: oauth_params["code_challenge"],
code_challenge_method: oauth_params["code_challenge_method"],
auth_time: Current.session.created_at.to_i,
acr: Current.session.acr,
expires_at: 10.minutes.from_now
)
# Clear OAuth params from session
session.delete(:oauth_params)
# Redirect back to client with authorization code
redirect_uri = "#{oauth_params['redirect_uri']}?code=#{code}"
redirect_uri += "&state=#{oauth_params['state']}" if oauth_params['state']
# Redirect back to client with authorization code (plaintext)
redirect_uri = "#{oauth_params["redirect_uri"]}?code=#{auth_code.plaintext_code}"
redirect_uri += "&state=#{CGI.escape(oauth_params["state"])}" if oauth_params["state"]
redirect_to redirect_uri, allow_other_host: true
end
@@ -261,24 +286,42 @@ class OidcController < ApplicationController
when "refresh_token"
handle_refresh_token_grant
else
render json: { error: "unsupported_grant_type" }, status: :bad_request
render json: {error: "unsupported_grant_type"}, status: :bad_request
end
end
def handle_authorization_code_grant
# Get client credentials from Authorization header or params
client_id, client_secret = extract_client_credentials
unless client_id && client_secret
render json: { error: "invalid_client" }, status: :unauthorized
unless client_id
render json: {error: "invalid_client", error_description: "client_id is required"}, status: :unauthorized
return
end
# Find and validate the application
# Find the application
application = Application.find_by(client_id: client_id)
unless application && application.authenticate_client_secret(client_secret)
render json: { error: "invalid_client" }, status: :unauthorized
unless application
render json: {error: "invalid_client", error_description: "Unknown client"}, status: :unauthorized
return
end
# Validate client credentials based on client type
if application.public_client?
# Public clients don't have a secret - they MUST use PKCE (checked later)
Rails.logger.info "OAuth: Public client authentication for #{application.name}"
else
# Confidential clients MUST provide valid client_secret
unless client_secret.present? && application.authenticate_client_secret(client_secret)
render json: {error: "invalid_client", error_description: "Invalid client credentials"}, status: :unauthorized
return
end
end
# Check if application is active
unless application.active?
Rails.logger.error "OAuth: Token request for inactive application: #{application.name}"
render json: {error: "invalid_client", error_description: "Application is not active"}, status: :forbidden
return
end
@@ -287,13 +330,11 @@ class OidcController < ApplicationController
redirect_uri = params[:redirect_uri]
code_verifier = params[:code_verifier]
auth_code = OidcAuthorizationCode.find_by(
application: application,
code: code
)
# Find authorization code using HMAC verification
auth_code = OidcAuthorizationCode.find_by_plaintext(code)
unless auth_code
render json: { error: "invalid_grant" }, status: :bad_request
unless auth_code && auth_code.application == application
render json: {error: "invalid_grant"}, status: :bad_request
return
end
@@ -324,18 +365,18 @@ class OidcController < ApplicationController
# Check if code is expired
if auth_code.expires_at < Time.current
render json: { error: "invalid_grant", error_description: "Authorization code expired" }, status: :bad_request
render json: {error: "invalid_grant", error_description: "Authorization code expired"}, status: :bad_request
return
end
# Validate redirect URI matches
unless auth_code.redirect_uri == redirect_uri
render json: { error: "invalid_grant", error_description: "Redirect URI mismatch" }, status: :bad_request
render json: {error: "invalid_grant", error_description: "Redirect URI mismatch"}, status: :bad_request
return
end
# Validate PKCE if code challenge is present
pkce_result = validate_pkce(auth_code, code_verifier)
# Validate PKCE - required for public clients and optionally for confidential clients
pkce_result = validate_pkce(application, auth_code, code_verifier)
unless pkce_result[:valid]
render json: {
error: pkce_result[:error],
@@ -362,7 +403,9 @@ class OidcController < ApplicationController
application: application,
user: user,
oidc_access_token: access_token_record,
scope: auth_code.scope
scope: auth_code.scope,
auth_time: auth_code.auth_time,
acr: auth_code.acr
)
# Find user consent for this application
@@ -370,12 +413,23 @@ class OidcController < ApplicationController
unless consent
Rails.logger.error "OIDC Security: Token requested without consent record (user: #{user.id}, app: #{application.id})"
render json: { error: "invalid_grant", error_description: "Authorization consent not found" }, status: :bad_request
render json: {error: "invalid_grant", error_description: "Authorization consent not found"}, status: :bad_request
return
end
# Generate ID token (JWT) with pairwise SID
id_token = OidcJwtService.generate_id_token(user, application, consent: consent, nonce: auth_code.nonce)
# Generate ID token (JWT) with pairwise SID, at_hash, auth_time, and acr
# auth_time and acr come from the authorization code (captured at /authorize time)
# scopes determine which claims are included (per OIDC Core spec)
id_token = OidcJwtService.generate_id_token(
user,
application,
consent: consent,
nonce: auth_code.nonce,
access_token: access_token_record.plaintext_token,
auth_time: auth_code.auth_time,
acr: auth_code.acr,
scopes: auth_code.scope
)
# Return tokens
render json: {
@@ -388,7 +442,7 @@ class OidcController < ApplicationController
}
end
rescue ActiveRecord::RecordNotFound
render json: { error: "invalid_grant" }, status: :bad_request
render json: {error: "invalid_grant"}, status: :bad_request
end
end
@@ -396,40 +450,56 @@ class OidcController < ApplicationController
# Get client credentials from Authorization header or params
client_id, client_secret = extract_client_credentials
unless client_id && client_secret
render json: { error: "invalid_client" }, status: :unauthorized
unless client_id
render json: {error: "invalid_client", error_description: "client_id is required"}, status: :unauthorized
return
end
# Find and validate the application
# Find the application
application = Application.find_by(client_id: client_id)
unless application && application.authenticate_client_secret(client_secret)
render json: { error: "invalid_client" }, status: :unauthorized
unless application
render json: {error: "invalid_client", error_description: "Unknown client"}, status: :unauthorized
return
end
# Validate client credentials based on client type
if application.public_client?
# Public clients don't have a secret
Rails.logger.info "OAuth: Public client refresh token request for #{application.name}"
else
# Confidential clients MUST provide valid client_secret
unless client_secret.present? && application.authenticate_client_secret(client_secret)
render json: {error: "invalid_client", error_description: "Invalid client credentials"}, status: :unauthorized
return
end
end
# Check if application is active
unless application.active?
Rails.logger.error "OAuth: Refresh token request for inactive application: #{application.name}"
render json: {error: "invalid_client", error_description: "Application is not active"}, status: :forbidden
return
end
# Get the refresh token
refresh_token = params[:refresh_token]
unless refresh_token.present?
render json: { error: "invalid_request", error_description: "refresh_token is required" }, status: :bad_request
render json: {error: "invalid_request", error_description: "refresh_token is required"}, status: :bad_request
return
end
# Find the refresh token record
# Note: This is inefficient with BCrypt hashing, but necessary for security
# In production, consider adding a token prefix for faster lookup
refresh_token_record = OidcRefreshToken.where(application: application).find do |rt|
rt.token_matches?(refresh_token)
end
# Find the refresh token record using indexed token prefix lookup
refresh_token_record = OidcRefreshToken.find_by_token(refresh_token)
unless refresh_token_record
render json: { error: "invalid_grant", error_description: "Invalid refresh token" }, status: :bad_request
# Verify the token belongs to the correct application
unless refresh_token_record && refresh_token_record.application == application
render json: {error: "invalid_grant", error_description: "Invalid refresh token"}, status: :bad_request
return
end
# Check if refresh token is expired
if refresh_token_record.expired?
render json: { error: "invalid_grant", error_description: "Refresh token expired" }, status: :bad_request
render json: {error: "invalid_grant", error_description: "Refresh token expired"}, status: :bad_request
return
end
@@ -440,7 +510,7 @@ class OidcController < ApplicationController
Rails.logger.warn "OAuth Security: Revoked refresh token reuse detected for token family #{refresh_token_record.token_family_id}"
refresh_token_record.revoke_family!
render json: { error: "invalid_grant", error_description: "Refresh token has been revoked" }, status: :bad_request
render json: {error: "invalid_grant", error_description: "Refresh token has been revoked"}, status: :bad_request
return
end
@@ -463,7 +533,9 @@ class OidcController < ApplicationController
user: user,
oidc_access_token: new_access_token,
scope: refresh_token_record.scope,
token_family_id: refresh_token_record.token_family_id # Keep same family for rotation tracking
token_family_id: refresh_token_record.token_family_id, # Keep same family for rotation tracking
auth_time: refresh_token_record.auth_time, # Carry over original auth_time
acr: refresh_token_record.acr # Carry over original acr
)
# Find user consent for this application
@@ -471,12 +543,22 @@ class OidcController < ApplicationController
unless consent
Rails.logger.error "OIDC Security: Refresh token used without consent record (user: #{user.id}, app: #{application.id})"
render json: { error: "invalid_grant", error_description: "Authorization consent not found" }, status: :bad_request
render json: {error: "invalid_grant", error_description: "Authorization consent not found"}, status: :bad_request
return
end
# Generate new ID token (JWT with pairwise SID, no nonce for refresh grants)
id_token = OidcJwtService.generate_id_token(user, application, consent: consent)
# Generate new ID token (JWT with pairwise SID, at_hash, auth_time, acr; no nonce for refresh grants)
# auth_time and acr come from the original refresh token (carried over from initial auth)
# scopes determine which claims are included (per OIDC Core spec)
id_token = OidcJwtService.generate_id_token(
user,
application,
consent: consent,
access_token: new_access_token.plaintext_token,
auth_time: refresh_token_record.auth_time,
acr: refresh_token_record.acr,
scopes: refresh_token_record.scope
)
# Return new tokens
render json: {
@@ -488,7 +570,7 @@ class OidcController < ApplicationController
scope: refresh_token_record.scope
}
rescue ActiveRecord::RecordNotFound
render json: { error: "invalid_grant" }, status: :bad_request
render json: {error: "invalid_grant"}, status: :bad_request
end
# GET /oauth/userinfo
@@ -509,6 +591,13 @@ class OidcController < ApplicationController
return
end
# Check if application is active (immediate cutoff when app is disabled)
unless access_token.application&.active?
Rails.logger.warn "OAuth: Userinfo request for inactive application: #{access_token.application&.name}"
head :forbidden
return
end
# Get the user (with fresh data from database)
user = access_token.user
unless user
@@ -565,19 +654,26 @@ class OidcController < ApplicationController
# Find and validate the application
application = Application.find_by(client_id: client_id)
unless application && application.authenticate_client_secret(client_secret)
unless application&.authenticate_client_secret(client_secret)
Rails.logger.warn "OAuth: Token revocation attempted for invalid application: #{client_id}"
head :ok
return
end
# Check if application is active (RFC 7009: still return 200 OK for privacy)
unless application.active?
Rails.logger.warn "OAuth: Token revocation attempted for inactive application: #{application.name}"
head :ok
return
end
# Get the token to revoke
token = params[:token]
token_type_hint = params[:token_type_hint] # Optional hint: "access_token" or "refresh_token"
unless token.present?
# RFC 7009: Missing token parameter is an error
render json: { error: "invalid_request", error_description: "token parameter is required" }, status: :bad_request
render json: {error: "invalid_request", error_description: "token parameter is required"}, status: :bad_request
return
end
@@ -587,9 +683,7 @@ class OidcController < ApplicationController
if token_type_hint == "refresh_token" || token_type_hint.nil?
# Try to find as refresh token
refresh_token_record = OidcRefreshToken.where(application: application).find do |rt|
rt.token_matches?(token)
end
refresh_token_record = OidcRefreshToken.find_by_token(token)
if refresh_token_record
refresh_token_record.revoke!
@@ -600,14 +694,12 @@ class OidcController < ApplicationController
if !revoked && (token_type_hint == "access_token" || token_type_hint.nil?)
# Try to find as access token
access_token_record = OidcAccessToken.where(application: application).find do |at|
at.token_matches?(token)
end
access_token_record = OidcAccessToken.find_by_token(token)
if access_token_record
access_token_record.revoke!
Rails.logger.info "OAuth: Access token revoked for application #{application.name}"
revoked = true
true
end
end
@@ -621,12 +713,17 @@ class OidcController < ApplicationController
# OpenID Connect RP-Initiated Logout
# Handle id_token_hint and post_logout_redirect_uri parameters
id_token_hint = params[:id_token_hint]
params[:id_token_hint]
post_logout_redirect_uri = params[:post_logout_redirect_uri]
state = params[:state]
# If user is authenticated, log them out
if authenticated?
user = Current.session.user
# Send backchannel logout notifications to all connected applications
send_backchannel_logout_notifications(user)
# Invalidate the current session
Current.session&.destroy
reset_session
@@ -638,7 +735,7 @@ class OidcController < ApplicationController
if validated_uri
redirect_uri = validated_uri
redirect_uri += "?state=#{state}" if state.present?
redirect_uri += "?state=#{CGI.escape(state)}" if state.present?
redirect_to redirect_uri, allow_other_host: true
else
# Invalid redirect URI - log warning and go to default
@@ -653,11 +750,26 @@ class OidcController < ApplicationController
private
def validate_pkce(auth_code, code_verifier)
# Skip PKCE validation if no code challenge was stored (legacy clients)
return { valid: true } unless auth_code.code_challenge.present?
def validate_pkce(application, auth_code, code_verifier)
# Check if PKCE is required for this application
pkce_required = application.requires_pkce?
pkce_provided = auth_code.code_challenge.present?
# PKCE is required but no verifier provided
# If PKCE is required but wasn't provided during authorization
if pkce_required && !pkce_provided
client_type = application.public_client? ? "public clients" : "this application"
return {
valid: false,
error: "invalid_request",
error_description: "PKCE is required for #{client_type}. code_challenge must be provided during authorization.",
status: :bad_request
}
end
# Skip validation if no code challenge was stored (legacy clients without PKCE requirement)
return {valid: true} unless pkce_provided
# PKCE was provided during authorization but no verifier sent with token request
unless code_verifier.present?
return {
valid: false,
@@ -702,7 +814,7 @@ class OidcController < ApplicationController
}
end
{ valid: true }
{valid: true}
end
def extract_client_credentials
@@ -727,7 +839,7 @@ class OidcController < ApplicationController
return nil unless parsed_uri.is_a?(URI::HTTP) || parsed_uri.is_a?(URI::HTTPS)
# Only allow HTTPS in production
return nil if Rails.env.production? && parsed_uri.scheme != 'https'
return nil if Rails.env.production? && parsed_uri.scheme != "https"
# Check if URI matches any registered OIDC application's redirect URIs
# According to OIDC spec, post_logout_redirect_uri should be pre-registered
@@ -766,4 +878,26 @@ class OidcController < ApplicationController
false
end
end
def send_backchannel_logout_notifications(user)
# Find all active OIDC consents for this user
consents = OidcUserConsent.where(user: user).includes(:application)
consents.each do |consent|
# Skip if application doesn't support backchannel logout
next unless consent.application.supports_backchannel_logout?
# Enqueue background job to send logout notification
BackchannelLogoutJob.perform_later(
user_id: user.id,
application_id: consent.application.id,
consent_sid: consent.sid
)
end
Rails.logger.info "OidcController: Enqueued #{consents.count} backchannel logout notifications for user #{user.id}"
rescue => e
# Log error but don't block logout
Rails.logger.error "OidcController: Failed to enqueue backchannel logout: #{e.class} - #{e.message}"
end
end

View File

@@ -1,13 +1,14 @@
class PasswordsController < ApplicationController
allow_unauthenticated_access
before_action :set_user_by_token, only: %i[ edit update ]
before_action :set_user_by_token, only: %i[edit update]
rate_limit to: 10, within: 3.minutes, only: :create, with: -> { redirect_to new_password_path, alert: "Try again later." }
rate_limit to: 10, within: 10.minutes, only: :update, with: -> { redirect_to new_password_path, alert: "Too many attempts. Try again later." }
def new
end
def create
if user = User.find_by(email_address: params[:email_address])
if (user = User.find_by(email_address: params[:email_address]))
PasswordsMailer.reset(user).deliver_later
end
@@ -27,6 +28,7 @@ class PasswordsController < ApplicationController
end
private
def set_user_by_token
@user = User.find_by_token_for(:password_reset, params[:token])
redirect_to new_password_path, alert: "Password reset link is invalid or has expired." if @user.nil?

View File

@@ -19,13 +19,21 @@ class ProfilesController < ApplicationController
else
render :show, status: :unprocessable_entity
end
else
# Updating email
elsif params[:user][:email_address].present?
# Updating email - requires current password (security: prevents account takeover)
unless @user.authenticate(params[:user][:current_password])
@user.errors.add(:current_password, "is required to change email")
render :show, status: :unprocessable_entity
return
end
if @user.update(email_params)
redirect_to profile_path, notice: "Email updated successfully."
else
render :show, status: :unprocessable_entity
end
else
render :show, status: :unprocessable_entity
end
end

View File

@@ -1,22 +1,22 @@
class SessionsController < ApplicationController
allow_unauthenticated_access only: %i[ new create verify_totp webauthn_challenge webauthn_verify ]
allow_unauthenticated_access only: %i[new create verify_totp webauthn_challenge webauthn_verify]
rate_limit to: 20, within: 3.minutes, only: :create, with: -> { redirect_to signin_path, alert: "Too many attempts. Try again later." }
rate_limit to: 10, within: 3.minutes, only: :verify_totp, with: -> { redirect_to totp_verification_path, alert: "Too many attempts. Try again later." }
rate_limit to: 10, within: 3.minutes, only: [:webauthn_challenge, :webauthn_verify], with: -> { render json: { error: "Too many attempts. Try again later." }, status: :too_many_requests }
rate_limit to: 10, within: 3.minutes, only: [:webauthn_challenge, :webauthn_verify], with: -> { render json: {error: "Too many attempts. Try again later."}, status: :too_many_requests }
def new
# Redirect to signup if this is first run
if User.count.zero?
respond_to do |format|
format.html { redirect_to signup_path }
format.json { render json: { error: "No users exist. Please complete initial setup." }, status: :service_unavailable }
format.json { render json: {error: "No users exist. Please complete initial setup."}, status: :service_unavailable }
end
return
end
respond_to do |format|
format.html # render HTML login page
format.json { render json: { error: "Authentication required" }, status: :unauthorized }
format.json { render json: {error: "Authentication required"}, status: :unauthorized }
end
end
@@ -71,8 +71,8 @@ class SessionsController < ApplicationController
return
end
# Sign in successful
start_new_session_for user
# Sign in successful (password only)
start_new_session_for user, acr: "1"
redirect_to after_authentication_url, notice: "Signed in successfully.", allow_other_host: true
end
@@ -101,39 +101,45 @@ class SessionsController < ApplicationController
return
end
# Try TOTP verification first
# Try TOTP verification first (password + TOTP = 2FA)
if user.verify_totp(code)
session.delete(:pending_totp_user_id)
# Restore redirect URL if it was preserved
if session[:totp_redirect_url].present?
session[:return_to_after_authenticating] = session.delete(:totp_redirect_url)
end
start_new_session_for user
start_new_session_for user, acr: "2"
redirect_to after_authentication_url, notice: "Signed in successfully.", allow_other_host: true
return
end
# Try backup code verification
# Try backup code verification (password + backup code = 2FA)
if user.verify_backup_code(code)
session.delete(:pending_totp_user_id)
# Restore redirect URL if it was preserved
if session[:totp_redirect_url].present?
session[:return_to_after_authenticating] = session.delete(:totp_redirect_url)
end
start_new_session_for user
start_new_session_for user, acr: "2"
redirect_to after_authentication_url, notice: "Signed in successfully using backup code.", allow_other_host: true
return
end
# Invalid code
redirect_to totp_verification_path, alert: "Invalid verification code. Please try again."
return
nil
end
# Just render the form
end
def destroy
# Send backchannel logout notifications before terminating session
if authenticated?
user = Current.session.user
send_backchannel_logout_notifications(user)
end
terminate_session
redirect_to signin_path, status: :see_other, notice: "Signed out successfully."
end
@@ -149,14 +155,14 @@ class SessionsController < ApplicationController
email = params[:email]&.strip&.downcase
if email.blank?
render json: { error: "Email is required" }, status: :unprocessable_entity
render json: {error: "Email is required"}, status: :unprocessable_entity
return
end
user = User.find_by(email_address: email)
if user.nil? || !user.can_authenticate_with_webauthn?
render json: { error: "User not found or WebAuthn not available" }, status: :unprocessable_entity
render json: {error: "User not found or WebAuthn not available"}, status: :unprocessable_entity
return
end
@@ -185,10 +191,9 @@ class SessionsController < ApplicationController
session[:webauthn_challenge] = options.challenge
render json: options
rescue => e
Rails.logger.error "WebAuthn challenge generation error: #{e.message}"
render json: { error: "Failed to generate WebAuthn challenge" }, status: :internal_server_error
render json: {error: "Failed to generate WebAuthn challenge"}, status: :internal_server_error
end
end
@@ -196,21 +201,21 @@ class SessionsController < ApplicationController
# Get pending user from session
user_id = session[:pending_webauthn_user_id]
unless user_id
render json: { error: "Session expired. Please try again." }, status: :unprocessable_entity
render json: {error: "Session expired. Please try again."}, status: :unprocessable_entity
return
end
user = User.find_by(id: user_id)
unless user
session.delete(:pending_webauthn_user_id)
render json: { error: "Session expired. Please try again." }, status: :unprocessable_entity
render json: {error: "Session expired. Please try again."}, status: :unprocessable_entity
return
end
# Get the credential and assertion from params
credential_data = params[:credential]
if credential_data.blank?
render json: { error: "Credential data is required" }, status: :unprocessable_entity
render json: {error: "Credential data is required"}, status: :unprocessable_entity
return
end
@@ -218,7 +223,7 @@ class SessionsController < ApplicationController
challenge = session.delete(:webauthn_challenge)
if challenge.blank?
render json: { error: "Invalid or expired session" }, status: :unprocessable_entity
render json: {error: "Invalid or expired session"}, status: :unprocessable_entity
return
end
@@ -231,7 +236,7 @@ class SessionsController < ApplicationController
stored_credential = user.webauthn_credential_for(external_id)
if stored_credential.nil?
render json: { error: "Credential not found" }, status: :unprocessable_entity
render json: {error: "Credential not found"}, status: :unprocessable_entity
return
end
@@ -262,24 +267,23 @@ class SessionsController < ApplicationController
session[:return_to_after_authenticating] = session.delete(:webauthn_redirect_url)
end
# Create session
start_new_session_for user
# Create session (WebAuthn/passkey = phishing-resistant, ACR = "2")
start_new_session_for user, acr: "2"
render json: {
success: true,
redirect_to: after_authentication_url,
message: "Signed in successfully with passkey"
}
rescue WebAuthn::Error => e
Rails.logger.error "WebAuthn verification error: #{e.message}"
render json: { error: "Authentication failed: #{e.message}" }, status: :unprocessable_entity
render json: {error: "Authentication failed: #{e.message}"}, status: :unprocessable_entity
rescue JSON::ParserError => e
Rails.logger.error "WebAuthn JSON parsing error: #{e.message}"
render json: { error: "Invalid credential format" }, status: :unprocessable_entity
render json: {error: "Invalid credential format"}, status: :unprocessable_entity
rescue => e
Rails.logger.error "Unexpected WebAuthn verification error: #{e.class} - #{e.message}"
render json: { error: "An unexpected error occurred" }, status: :internal_server_error
render json: {error: "An unexpected error occurred"}, status: :internal_server_error
end
end
@@ -295,7 +299,7 @@ class SessionsController < ApplicationController
return nil unless uri.is_a?(URI::HTTP) || uri.is_a?(URI::HTTPS)
# Only allow HTTPS in production
return nil unless Rails.env.development? || uri.scheme == 'https'
return nil unless Rails.env.development? || uri.scheme == "https"
redirect_domain = uri.host.downcase
return nil unless redirect_domain.present?
@@ -306,9 +310,30 @@ class SessionsController < ApplicationController
end
matching_app ? url : nil
rescue URI::InvalidURIError
nil
end
end
def send_backchannel_logout_notifications(user)
# Find all active OIDC consents for this user
consents = OidcUserConsent.where(user: user).includes(:application)
consents.each do |consent|
# Skip if application doesn't support backchannel logout
next unless consent.application.supports_backchannel_logout?
# Enqueue background job to send logout notification
BackchannelLogoutJob.perform_later(
user_id: user.id,
application_id: consent.application.id,
consent_sid: consent.sid
)
end
Rails.logger.info "SessionsController: Enqueued #{consents.count} backchannel logout notifications for user #{user.id}"
rescue => e
# Log error but don't block logout
Rails.logger.error "SessionsController: Failed to enqueue backchannel logout: #{e.class} - #{e.message}"
end
end

View File

@@ -1,6 +1,6 @@
class UsersController < ApplicationController
allow_unauthenticated_access only: %i[ new create ]
before_action :ensure_first_run, only: %i[ new create ]
allow_unauthenticated_access only: %i[new create]
before_action :ensure_first_run, only: %i[new create]
def new
@user = User.new

View File

@@ -2,6 +2,11 @@ class WebauthnController < ApplicationController
before_action :set_webauthn_credential, only: [:destroy]
skip_before_action :require_authentication, only: [:check]
# Rate limit check endpoint to prevent enumeration attacks
rate_limit to: 10, within: 1.minute, only: [:check], with: -> {
render json: {error: "Too many requests. Try again later."}, status: :too_many_requests
}
# GET /webauthn/new
def new
@webauthn_credential = WebauthnCredential.new
@@ -11,7 +16,7 @@ class WebauthnController < ApplicationController
# Generate registration challenge for creating a new passkey
def challenge
user = Current.session&.user
return render json: { error: "Not authenticated" }, status: :unauthorized unless user
return render json: {error: "Not authenticated"}, status: :unauthorized unless user
registration_options = WebAuthn::Credential.options_for_create(
user: {
@@ -39,7 +44,7 @@ class WebauthnController < ApplicationController
credential_data, nickname = extract_credential_params
if credential_data.blank? || nickname.blank?
render json: { error: "Credential and nickname are required" }, status: :unprocessable_entity
render json: {error: "Credential and nickname are required"}, status: :unprocessable_entity
return
end
@@ -47,7 +52,7 @@ class WebauthnController < ApplicationController
challenge = session.delete(:webauthn_challenge)
if challenge.blank?
render json: { error: "Invalid or expired session" }, status: :unprocessable_entity
render json: {error: "Invalid or expired session"}, status: :unprocessable_entity
return
end
@@ -74,7 +79,7 @@ class WebauthnController < ApplicationController
# Store the credential
user = Current.session&.user
return render json: { error: "Not authenticated" }, status: :unauthorized unless user
return render json: {error: "Not authenticated"}, status: :unauthorized unless user
@webauthn_credential = user.webauthn_credentials.create!(
external_id: Base64.urlsafe_encode64(webauthn_credential.id),
@@ -91,27 +96,18 @@ class WebauthnController < ApplicationController
message: "Passkey '#{nickname}' registered successfully",
credential_id: @webauthn_credential.id
}
rescue WebAuthn::Error => e
Rails.logger.error "WebAuthn registration error: #{e.message}"
render json: { error: "Failed to register passkey: #{e.message}" }, status: :unprocessable_entity
render json: {error: "Failed to register passkey: #{e.message}"}, status: :unprocessable_entity
rescue => e
Rails.logger.error "Unexpected WebAuthn registration error: #{e.class} - #{e.message}"
render json: { error: "An unexpected error occurred" }, status: :internal_server_error
render json: {error: "An unexpected error occurred"}, status: :internal_server_error
end
end
# DELETE /webauthn/:id
# Remove a passkey
def destroy
user = Current.session&.user
return render json: { error: "Not authenticated" }, status: :unauthorized unless user
if @webauthn_credential.user != user
render json: { error: "Unauthorized" }, status: :forbidden
return
end
nickname = @webauthn_credential.nickname
@webauthn_credential.destroy
@@ -131,25 +127,27 @@ class WebauthnController < ApplicationController
# GET /webauthn/check
# Check if user has WebAuthn credentials (for login page detection)
# Security: Returns identical responses for non-existent users to prevent enumeration
def check
email = params[:email]&.strip&.downcase
if email.blank?
render json: { has_webauthn: false, error: "Email is required" }
render json: {has_webauthn: false, requires_webauthn: false}
return
end
user = User.find_by(email_address: email)
# Security: Return identical response for non-existent users
# Combined with rate limiting (10/min), this prevents account enumeration
if user.nil?
render json: { has_webauthn: false, message: "User not found" }
render json: {has_webauthn: false, requires_webauthn: false}
return
end
# Only return minimal necessary info - no user_id or preferred_method
render json: {
has_webauthn: user.can_authenticate_with_webauthn?,
user_id: user.id,
preferred_method: user.preferred_authentication_method,
requires_webauthn: user.require_webauthn?
}
end
@@ -159,7 +157,7 @@ class WebauthnController < ApplicationController
def extract_credential_params
# Use require.permit which is working and reliable
# The JavaScript sends params both directly and wrapped in webauthn key
begin
# Try direct parameters first
credential_params = params.require(:credential).permit(:id, :rawId, :type, response: {}, clientExtensionResults: {})
nickname = params.require(:nickname)
@@ -170,29 +168,25 @@ class WebauthnController < ApplicationController
webauthn_params = params.require(:webauthn).permit(:nickname, credential: [:id, :rawId, :type, response: {}, clientExtensionResults: {}])
[webauthn_params[:credential], webauthn_params[:nickname]]
end
end
def set_webauthn_credential
@webauthn_credential = WebauthnCredential.find(params[:id])
user = Current.session&.user
return render json: {error: "Not authenticated"}, status: :unauthorized unless user
@webauthn_credential = user.webauthn_credentials.find(params[:id])
rescue ActiveRecord::RecordNotFound
respond_to do |format|
format.html {
redirect_to profile_path,
alert: "Passkey not found"
}
format.json {
render json: { error: "Passkey not found" }, status: :not_found
}
format.html { redirect_to profile_path, alert: "Passkey not found" }
format.json { render json: {error: "Passkey not found"}, status: :not_found }
end
end
# Helper method to convert Base64 to Base64URL if needed
def base64_to_base64url(str)
str.gsub('+', '-').gsub('/', '_').gsub(/=+$/, '')
str.tr("+", "-").tr("/", "_").gsub(/=+$/, "")
end
# Helper method to convert Base64URL to Base64 if needed
def base64url_to_base64(str)
str.gsub('-', '+').gsub('_', '/') + '=' * (4 - str.length % 4) % 4
str.tr("-", "+").tr("_", "/") + "=" * (4 - str.length % 4) % 4
end
end

View File

@@ -22,11 +22,11 @@ module ApplicationHelper
def border_class_for(type)
case type.to_s
when 'notice' then 'border-green-200'
when 'alert', 'error' then 'border-red-200'
when 'warning' then 'border-yellow-200'
when 'info' then 'border-blue-200'
else 'border-gray-200'
when "notice" then "border-green-200"
when "alert", "error" then "border-red-200"
when "warning" then "border-yellow-200"
when "info" then "border-blue-200"
else "border-gray-200"
end
end
end

View File

@@ -25,9 +25,7 @@ module ClaimsHelper
claims = deep_merge_claims(claims, user.parsed_custom_claims)
# Merge app-specific claims (arrays are combined)
claims = deep_merge_claims(claims, application.custom_claims_for_user(user))
claims
deep_merge_claims(claims, application.custom_claims_for_user(user))
end
# Get claim sources breakdown for display

View File

@@ -1,7 +1,7 @@
import { Controller } from "@hotwired/stimulus"
export default class extends Controller {
static targets = ["appTypeSelect", "oidcFields", "forwardAuthFields"]
static targets = ["appTypeSelect", "oidcFields", "forwardAuthFields", "pkceOptions"]
connect() {
this.updateFieldVisibility()
@@ -21,4 +21,17 @@ export default class extends Controller {
this.forwardAuthFieldsTarget.classList.add('hidden')
}
}
updatePkceVisibility(event) {
// Show PKCE options for confidential clients, hide for public clients
const isPublicClient = event.target.value === "true"
if (this.hasPkceOptionsTarget) {
if (isPublicClient) {
this.pkceOptionsTarget.classList.add('hidden')
} else {
this.pkceOptionsTarget.classList.remove('hidden')
}
}
}
}

View File

@@ -0,0 +1,121 @@
import { Controller } from "@hotwired/stimulus"
export default class extends Controller {
static targets = ["input", "dropzone"]
connect() {
// Listen for paste events on the dropzone
this.dropzoneTarget.addEventListener("paste", this.handlePaste.bind(this))
}
disconnect() {
this.dropzoneTarget.removeEventListener("paste", this.handlePaste.bind(this))
}
handlePaste(e) {
e.preventDefault()
e.stopPropagation()
const clipboardData = e.clipboardData || e.originalEvent.clipboardData
// First, try to get image data
for (let item of clipboardData.items) {
if (item.type.indexOf("image") !== -1) {
const blob = item.getAsFile()
this.handleImageBlob(blob)
return
}
}
// If no image found, check for SVG text
const text = clipboardData.getData("text/plain")
if (text && this.isSVG(text)) {
this.handleSVGText(text)
return
}
}
isSVG(text) {
// Check if the text looks like SVG code
const trimmed = text.trim()
return trimmed.startsWith("<svg") && trimmed.includes("</svg>")
}
handleSVGText(svgText) {
// Validate file size (2MB)
const size = new Blob([svgText]).size
if (size > 2 * 1024 * 1024) {
alert("SVG code is too large (must be less than 2MB)")
return
}
// Create a blob from the SVG text
const blob = new Blob([svgText], { type: "image/svg+xml" })
// Create a File object
const file = new File([blob], `pasted-svg-${Date.now()}.svg`, {
type: "image/svg+xml"
})
// Create a DataTransfer object to set files on the input
const dataTransfer = new DataTransfer()
dataTransfer.items.add(file)
this.inputTarget.files = dataTransfer.files
// Trigger change event to update preview (file-drop controller will handle it)
const event = new Event("change", { bubbles: true })
this.inputTarget.dispatchEvent(event)
// Visual feedback
this.dropzoneTarget.classList.add("border-green-500", "bg-green-50")
setTimeout(() => {
this.dropzoneTarget.classList.remove("border-green-500", "bg-green-50")
}, 500)
}
handleImageBlob(blob) {
// Validate file type
const validTypes = ["image/png", "image/jpg", "image/jpeg", "image/gif", "image/svg+xml"]
if (!validTypes.includes(blob.type)) {
alert("Please paste a PNG, JPG, GIF, or SVG image")
return
}
// Validate file size (2MB)
if (blob.size > 2 * 1024 * 1024) {
alert("Image size must be less than 2MB")
return
}
// Create a File object from the blob with a default name
const file = new File([blob], `pasted-image-${Date.now()}.${this.getExtension(blob.type)}`, {
type: blob.type
})
// Create a DataTransfer object to set files on the input
const dataTransfer = new DataTransfer()
dataTransfer.items.add(file)
this.inputTarget.files = dataTransfer.files
// Trigger change event to update preview (file-drop controller will handle it)
const event = new Event("change", { bubbles: true })
this.inputTarget.dispatchEvent(event)
// Visual feedback
this.dropzoneTarget.classList.add("border-green-500", "bg-green-50")
setTimeout(() => {
this.dropzoneTarget.classList.remove("border-green-500", "bg-green-50")
}, 500)
}
getExtension(mimeType) {
const extensions = {
"image/png": "png",
"image/jpeg": "jpg",
"image/jpg": "jpg",
"image/gif": "gif",
"image/svg+xml": "svg"
}
return extensions[mimeType] || "png"
}
}

View File

@@ -0,0 +1,52 @@
class BackchannelLogoutJob < ApplicationJob
queue_as :default
# Retry with exponential backoff: 1s, 5s, 25s
retry_on StandardError, wait: :exponentially_longer, attempts: 3
def perform(user_id:, application_id:, consent_sid:)
# Find the records
user = User.find_by(id: user_id)
application = Application.find_by(id: application_id)
consent = OidcUserConsent.find_by(sid: consent_sid)
# Validate we have all required data
unless user && application && consent
Rails.logger.warn "BackchannelLogout: Missing data - user: #{user.present?}, app: #{application.present?}, consent: #{consent.present?}"
return
end
# Skip if application doesn't support backchannel logout
unless application.supports_backchannel_logout?
Rails.logger.debug "BackchannelLogout: Application #{application.name} doesn't support backchannel logout"
return
end
# Generate the logout token
logout_token = OidcJwtService.generate_logout_token(user, application, consent)
# Send HTTP POST to the application's backchannel logout URI
uri = URI.parse(application.backchannel_logout_uri)
begin
response = Net::HTTP.start(uri.host, uri.port, use_ssl: uri.scheme == "https", open_timeout: 5, read_timeout: 5) do |http|
request = Net::HTTP::Post.new(uri.path.presence || "/")
request["Content-Type"] = "application/x-www-form-urlencoded"
request.set_form_data({logout_token: logout_token})
http.request(request)
end
if response.code.to_i == 200
Rails.logger.info "BackchannelLogout: Successfully sent logout notification to #{application.name} (#{application.backchannel_logout_uri})"
else
Rails.logger.warn "BackchannelLogout: Application #{application.name} returned HTTP #{response.code} from #{application.backchannel_logout_uri}"
end
rescue Net::OpenTimeout, Net::ReadTimeout => e
Rails.logger.warn "BackchannelLogout: Timeout sending logout to #{application.name} (#{application.backchannel_logout_uri}): #{e.message}"
raise # Retry on timeout
rescue => e
Rails.logger.error "BackchannelLogout: Failed to send logout to #{application.name} (#{application.backchannel_logout_uri}): #{e.class} - #{e.message}"
raise # Retry on error
end
end
end

View File

@@ -1,4 +1,4 @@
class ApplicationMailer < ActionMailer::Base
default from: ENV.fetch('CLINCH_FROM_EMAIL', 'clinch@example.com')
default from: ENV.fetch("CLINCH_FROM_EMAIL", "clinch@example.com")
layout "mailer"
end

View File

@@ -1,6 +1,15 @@
class Application < ApplicationRecord
has_secure_password :client_secret, validations: false
# Virtual attribute to control client type during creation
# When true, no client_secret will be generated (public client)
attr_accessor :is_public_client
has_one_attached :icon
# Fix SVG content type after attachment
after_save :fix_icon_content_type, if: -> { icon.attached? && saved_change_to_attribute?(:id) == false }
has_many :application_groups, dependent: :destroy
has_many :allowed_groups, through: :application_groups, source: :group
has_many :application_user_claims, dependent: :destroy
@@ -10,35 +19,48 @@ class Application < ApplicationRecord
has_many :oidc_user_consents, dependent: :destroy
validates :name, presence: true
validates :slug, presence: true, uniqueness: { case_sensitive: false },
format: { with: /\A[a-z0-9\-]+\z/, message: "only lowercase letters, numbers, and hyphens" }
validates :slug, presence: true, uniqueness: {case_sensitive: false},
format: {with: /\A[a-z0-9-]+\z/, message: "only lowercase letters, numbers, and hyphens"}
validates :app_type, presence: true,
inclusion: { in: %w[oidc forward_auth] }
validates :client_id, uniqueness: { allow_nil: true }
validates :client_secret, presence: true, on: :create, if: -> { oidc? }
validates :domain_pattern, presence: true, uniqueness: { case_sensitive: false }, if: :forward_auth?
validates :landing_url, format: { with: URI::regexp(%w[http https]), allow_nil: true, message: "must be a valid URL" }
inclusion: {in: %w[oidc forward_auth]}
validates :client_id, uniqueness: {allow_nil: true}
validates :client_secret, presence: true, on: :create, if: -> { oidc? && confidential_client? }
validates :domain_pattern, presence: true, uniqueness: {case_sensitive: false}, if: :forward_auth?
validates :landing_url, format: {with: URI::RFC2396_PARSER.make_regexp(%w[http https]), allow_nil: true, message: "must be a valid URL"}
validates :backchannel_logout_uri, format: {
with: URI::RFC2396_PARSER.make_regexp(%w[http https]),
allow_nil: true,
message: "must be a valid HTTP or HTTPS URL"
}
validate :backchannel_logout_uri_must_be_https_in_production, if: -> { backchannel_logout_uri.present? }
# Icon validation using ActiveStorage validators
validate :icon_validation, if: -> { icon.attached? }
# Token TTL validations (for OIDC apps)
validates :access_token_ttl, numericality: { greater_than_or_equal_to: 300, less_than_or_equal_to: 86400 }, if: :oidc? # 5 min - 24 hours
validates :refresh_token_ttl, numericality: { greater_than_or_equal_to: 86400, less_than_or_equal_to: 7776000 }, if: :oidc? # 1 day - 90 days
validates :id_token_ttl, numericality: { greater_than_or_equal_to: 300, less_than_or_equal_to: 86400 }, if: :oidc? # 5 min - 24 hours
validates :access_token_ttl, numericality: {greater_than_or_equal_to: 300, less_than_or_equal_to: 86400}, if: :oidc? # 5 min - 24 hours
validates :refresh_token_ttl, numericality: {greater_than_or_equal_to: 86400, less_than_or_equal_to: 7776000}, if: :oidc? # 1 day - 90 days
validates :id_token_ttl, numericality: {greater_than_or_equal_to: 300, less_than_or_equal_to: 86400}, if: :oidc? # 5 min - 24 hours
normalizes :slug, with: ->(slug) { slug.strip.downcase }
normalizes :domain_pattern, with: ->(pattern) {
normalized = pattern&.strip&.downcase
normalized.blank? ? nil : normalized
}
normalizes :backchannel_logout_uri, with: ->(uri) {
normalized = uri&.strip
normalized.blank? ? nil : normalized
}
before_validation :generate_client_credentials, on: :create, if: :oidc?
# Default header configuration for ForwardAuth
DEFAULT_HEADERS = {
user: 'X-Remote-User',
email: 'X-Remote-Email',
name: 'X-Remote-Name',
groups: 'X-Remote-Groups',
admin: 'X-Remote-Admin'
user: "X-Remote-User",
email: "X-Remote-Email",
name: "X-Remote-Name",
groups: "X-Remote-Groups",
admin: "X-Remote-Admin"
}.freeze
# Scopes
@@ -56,6 +78,24 @@ class Application < ApplicationRecord
app_type == "forward_auth"
end
# Client type checks (for OIDC)
def public_client?
client_secret_digest.blank?
end
def confidential_client?
!public_client?
end
# PKCE requirement check
# Public clients MUST use PKCE (no client secret to protect auth code)
# Confidential clients can optionally require PKCE (OAuth 2.1 recommendation)
def requires_pkce?
return false unless oidc?
return true if public_client? # Always require PKCE for public clients
require_pkce? # Check the flag for confidential clients
end
# Access control
def user_allowed?(user)
return false unless active?
@@ -95,8 +135,8 @@ class Application < ApplicationRecord
def matches_domain?(domain)
return false if domain.blank? || !forward_auth?
pattern = domain_pattern.gsub('.', '\.')
pattern = pattern.gsub('*', '[^.]*')
pattern = domain_pattern.gsub(".", '\.')
pattern = pattern.gsub("*", "[^.]*")
regex = Regexp.new("^#{pattern}$", Regexp::IGNORECASE)
regex.match?(domain.downcase)
@@ -104,18 +144,18 @@ class Application < ApplicationRecord
# Policy determination based on user status (for ForwardAuth)
def policy_for_user(user)
return 'deny' unless active?
return 'deny' unless user.active?
return "deny" unless active?
return "deny" unless user.active?
# If no groups specified, bypass authentication
return 'bypass' if allowed_groups.empty?
return "bypass" if allowed_groups.empty?
# If user is in allowed groups, determine auth level
if user_allowed?(user)
# Require 2FA if user has TOTP configured, otherwise one factor
user.totp_enabled? ? 'two_factor' : 'one_factor'
user.totp_enabled? ? "two_factor" : "one_factor"
else
'deny'
"deny"
end
end
@@ -157,7 +197,7 @@ class Application < ApplicationRecord
def generate_new_client_secret!
secret = SecureRandom.urlsafe_base64(48)
self.client_secret = secret
self.save!
save!
secret
end
@@ -193,8 +233,44 @@ class Application < ApplicationRecord
app_claim&.parsed_custom_claims || {}
end
# Check if this application supports backchannel logout
def supports_backchannel_logout?
backchannel_logout_uri.present?
end
# Check if a user has an active session with this application
# (i.e., has valid, non-revoked tokens)
def user_has_active_session?(user)
oidc_access_tokens.where(user: user).valid.exists? ||
oidc_refresh_tokens.where(user: user).valid.exists?
end
private
def fix_icon_content_type
return unless icon.attached?
# Fix SVG content type if it was detected incorrectly
if icon.filename.extension == "svg" && icon.content_type == "application/octet-stream"
icon.blob.update(content_type: "image/svg+xml")
end
end
def icon_validation
return unless icon.attached?
# Check content type
allowed_types = ["image/png", "image/jpg", "image/jpeg", "image/gif", "image/svg+xml"]
unless allowed_types.include?(icon.content_type)
errors.add(:icon, "must be a PNG, JPG, GIF, or SVG image")
end
# Check file size (2MB limit)
if icon.blob.byte_size > 2.megabytes
errors.add(:icon, "must be less than 2MB")
end
end
def duration_to_human(seconds)
if seconds < 3600
"#{seconds / 60} minutes"
@@ -207,10 +283,30 @@ class Application < ApplicationRecord
def generate_client_credentials
self.client_id ||= SecureRandom.urlsafe_base64(32)
# Generate and hash the client secret
if new_record? && client_secret.blank?
# Generate client secret only for confidential clients
# Public clients (is_public_client checked) don't get a secret - they use PKCE only
if new_record? && client_secret.blank? && !is_public_client_selected?
secret = SecureRandom.urlsafe_base64(48)
self.client_secret = secret
end
end
# Check if the user selected public client option
def is_public_client_selected?
ActiveModel::Type::Boolean.new.cast(is_public_client)
end
def backchannel_logout_uri_must_be_https_in_production
return unless Rails.env.production?
return unless backchannel_logout_uri.present?
begin
uri = URI.parse(backchannel_logout_uri)
unless uri.scheme == "https"
errors.add(:backchannel_logout_uri, "must use HTTPS in production")
end
rescue URI::InvalidURIError
# Let the format validator handle invalid URIs
end
end
end

View File

@@ -2,5 +2,5 @@ class ApplicationGroup < ApplicationRecord
belongs_to :application
belongs_to :group
validates :application_id, uniqueness: { scope: :group_id }
validates :application_id, uniqueness: {scope: :group_id}
end

View File

@@ -9,7 +9,7 @@ class ApplicationUserClaim < ApplicationRecord
groups
].freeze
validates :user_id, uniqueness: { scope: :application_id }
validates :user_id, uniqueness: {scope: :application_id}
validate :no_reserved_claim_names
# Parse custom_claims JSON field
@@ -25,7 +25,7 @@ class ApplicationUserClaim < ApplicationRecord
reserved_used = parsed_custom_claims.keys.map(&:to_s) & RESERVED_CLAIMS
if reserved_used.any?
errors.add(:custom_claims, "cannot override reserved OIDC claims: #{reserved_used.join(', ')}")
errors.add(:custom_claims, "cannot override reserved OIDC claims: #{reserved_used.join(", ")}")
end
end
end

View File

@@ -11,7 +11,7 @@ class Group < ApplicationRecord
groups
].freeze
validates :name, presence: true, uniqueness: { case_sensitive: false }
validates :name, presence: true, uniqueness: {case_sensitive: false}
normalizes :name, with: ->(name) { name.strip.downcase }
validate :no_reserved_claim_names
@@ -28,7 +28,7 @@ class Group < ApplicationRecord
reserved_used = parsed_custom_claims.keys.map(&:to_s) & RESERVED_CLAIMS
if reserved_used.any?
errors.add(:custom_claims, "cannot override reserved OIDC claims: #{reserved_used.join(', ')}")
errors.add(:custom_claims, "cannot override reserved OIDC claims: #{reserved_used.join(", ")}")
end
end
end

View File

@@ -6,7 +6,7 @@ class OidcAccessToken < ApplicationRecord
before_validation :generate_token, on: :create
before_validation :set_expiry, on: :create
validates :token, uniqueness: true, presence: true
validates :token_hmac, presence: true, uniqueness: true
scope :valid, -> { where("expires_at > ?", Time.current).where(revoked_at: nil) }
scope :expired, -> { where("expires_at <= ?", Time.current) }
@@ -15,6 +15,19 @@ class OidcAccessToken < ApplicationRecord
attr_accessor :plaintext_token # Store plaintext temporarily for returning to client
# Find access token by plaintext token using HMAC verification
def self.find_by_token(plaintext_token)
return nil if plaintext_token.blank?
token_hmac = compute_token_hmac(plaintext_token)
find_by(token_hmac: token_hmac)
end
# Compute HMAC for token lookup
def self.compute_token_hmac(plaintext_token)
OpenSSL::HMAC.hexdigest("SHA256", TokenHmac::KEY, plaintext_token)
end
def expired?
expires_at <= Time.current
end
@@ -33,48 +46,13 @@ class OidcAccessToken < ApplicationRecord
oidc_refresh_tokens.each(&:revoke!)
end
# Check if a plaintext token matches the hashed token
def token_matches?(plaintext_token)
return false if plaintext_token.blank?
# Use BCrypt to compare if token_digest exists
if token_digest.present?
BCrypt::Password.new(token_digest) == plaintext_token
# Fall back to direct comparison for backward compatibility
elsif token.present?
token == plaintext_token
else
false
end
end
# Find by token (validates and checks if revoked)
def self.find_by_token(plaintext_token)
return nil if plaintext_token.blank?
# Find all non-revoked, non-expired tokens
valid.find_each do |access_token|
# Use BCrypt to compare (if token_digest exists) or direct comparison
if access_token.token_digest.present?
return access_token if BCrypt::Password.new(access_token.token_digest) == plaintext_token
elsif access_token.token == plaintext_token
return access_token
end
end
nil
end
private
def generate_token
return if token.present?
# Generate opaque access token
plaintext = SecureRandom.urlsafe_base64(48)
self.plaintext_token = plaintext # Store temporarily for returning to client
self.token_digest = BCrypt::Password.create(plaintext)
# Keep token column for backward compatibility during migration
self.token = plaintext
# Generate random plaintext token
self.plaintext_token ||= SecureRandom.urlsafe_base64(48)
# Store HMAC in database (not plaintext)
self.token_hmac ||= self.class.compute_token_hmac(plaintext_token)
end
def set_expiry

View File

@@ -2,17 +2,32 @@ class OidcAuthorizationCode < ApplicationRecord
belongs_to :application
belongs_to :user
attr_accessor :plaintext_code
before_validation :generate_code, on: :create
before_validation :set_expiry, on: :create
validates :code, presence: true, uniqueness: true
validates :code_hmac, presence: true, uniqueness: true
validates :redirect_uri, presence: true
validates :code_challenge_method, inclusion: { in: %w[plain S256], allow_nil: true }
validates :code_challenge_method, inclusion: {in: %w[plain S256], allow_nil: true}
validate :validate_code_challenge_format, if: -> { code_challenge.present? }
scope :valid, -> { where(used: false).where("expires_at > ?", Time.current) }
scope :expired, -> { where("expires_at <= ?", Time.current) }
# Find authorization code by plaintext code using HMAC verification
def self.find_by_plaintext(plaintext_code)
return nil if plaintext_code.blank?
code_hmac = compute_code_hmac(plaintext_code)
find_by(code_hmac: code_hmac)
end
# Compute HMAC for code lookup
def self.compute_code_hmac(plaintext_code)
OpenSSL::HMAC.hexdigest("SHA256", TokenHmac::KEY, plaintext_code)
end
def expired?
expires_at <= Time.current
end
@@ -32,7 +47,10 @@ class OidcAuthorizationCode < ApplicationRecord
private
def generate_code
self.code ||= SecureRandom.urlsafe_base64(32)
# Generate random plaintext code
self.plaintext_code ||= SecureRandom.urlsafe_base64(32)
# Store HMAC in database (not plaintext)
self.code_hmac ||= self.class.compute_code_hmac(plaintext_code)
end
def set_expiry

View File

@@ -2,13 +2,12 @@ class OidcRefreshToken < ApplicationRecord
belongs_to :application
belongs_to :user
belongs_to :oidc_access_token
has_many :oidc_access_tokens, foreign_key: :oidc_access_token_id, dependent: :nullify
before_validation :generate_token, on: :create
before_validation :set_expiry, on: :create
before_validation :set_token_family_id, on: :create
validates :token_digest, presence: true, uniqueness: true
validates :token_hmac, presence: true, uniqueness: true
scope :valid, -> { where("expires_at > ?", Time.current).where(revoked_at: nil) }
scope :expired, -> { where("expires_at <= ?", Time.current) }
@@ -20,6 +19,19 @@ class OidcRefreshToken < ApplicationRecord
attr_accessor :token # Store plaintext token temporarily for returning to client
# Find refresh token by plaintext token using HMAC verification
def self.find_by_token(plaintext_token)
return nil if plaintext_token.blank?
token_hmac = compute_token_hmac(plaintext_token)
find_by(token_hmac: token_hmac)
end
# Compute HMAC for token lookup
def self.compute_token_hmac(plaintext_token)
OpenSSL::HMAC.hexdigest("SHA256", TokenHmac::KEY, plaintext_token)
end
def expired?
expires_at <= Time.current
end
@@ -43,35 +55,13 @@ class OidcRefreshToken < ApplicationRecord
OidcRefreshToken.in_family(token_family_id).update_all(revoked_at: Time.current)
end
# Verify a plaintext token against the stored digest
def self.find_by_token(plaintext_token)
return nil if plaintext_token.blank?
# Try to find tokens that could match (we can't search by hash directly)
# This is less efficient but necessary with BCrypt
# In production, you might want to add a token prefix or other optimization
all.find do |refresh_token|
refresh_token.token_matches?(plaintext_token)
end
end
def token_matches?(plaintext_token)
return false if plaintext_token.blank? || token_digest.blank?
BCrypt::Password.new(token_digest) == plaintext_token
rescue BCrypt::Errors::InvalidHash
false
end
private
def generate_token
# Generate a secure random token
plaintext = SecureRandom.urlsafe_base64(48)
self.token = plaintext # Store temporarily for returning to client
# Hash it with BCrypt for storage
self.token_digest = BCrypt::Password.create(plaintext)
# Generate random plaintext token
self.token ||= SecureRandom.urlsafe_base64(48)
# Store HMAC in database (not plaintext)
self.token_hmac ||= self.class.compute_token_hmac(token)
end
def set_expiry

View File

@@ -3,19 +3,19 @@ class OidcUserConsent < ApplicationRecord
belongs_to :application
validates :user, :application, :scopes_granted, :granted_at, presence: true
validates :user_id, uniqueness: { scope: :application_id }
validates :user_id, uniqueness: {scope: :application_id}
before_validation :set_granted_at, on: :create
before_validation :set_sid, on: :create
# Parse scopes_granted into an array
def scopes
scopes_granted.split(' ')
scopes_granted.split(" ")
end
# Set scopes from an array
def scopes=(scope_array)
self.scopes_granted = Array(scope_array).uniq.join(' ')
self.scopes_granted = Array(scope_array).uniq.join(" ")
end
# Check if this consent covers the requested scopes
@@ -31,18 +31,18 @@ class OidcUserConsent < ApplicationRecord
def formatted_scopes
scopes.map do |scope|
case scope
when 'openid'
'Basic authentication'
when 'profile'
'Profile information'
when 'email'
'Email address'
when 'groups'
'Group membership'
when "openid"
"Basic authentication"
when "profile"
"Profile information"
when "email"
"Email address"
when "groups"
"Group membership"
else
scope.humanize
end
end.join(', ')
end.join(", ")
end
# Find consent by SID

View File

@@ -1,4 +1,7 @@
class User < ApplicationRecord
# Encrypt TOTP secrets at rest (key derived from SECRET_KEY_BASE)
encrypts :totp_secret
has_secure_password
has_many :sessions, dependent: :destroy
has_many :user_groups, dependent: :destroy
@@ -16,10 +19,6 @@ class User < ApplicationRecord
updated_at
end
generates_token_for :magic_login, expires_in: 15.minutes do
last_sign_in_at
end
normalizes :email_address, with: ->(e) { e.strip.downcase }
normalizes :username, with: ->(u) { u.strip.downcase if u.present? }
@@ -30,16 +29,16 @@ class User < ApplicationRecord
groups
].freeze
validates :email_address, presence: true, uniqueness: { case_sensitive: false },
format: { with: URI::MailTo::EMAIL_REGEXP }
validates :username, uniqueness: { case_sensitive: false }, allow_nil: true,
format: { with: /\A[a-zA-Z0-9_-]+\z/, message: "can only contain letters, numbers, underscores, and hyphens" },
length: { minimum: 2, maximum: 30 }
validates :password, length: { minimum: 8 }, allow_nil: true
validates :email_address, presence: true, uniqueness: {case_sensitive: false},
format: {with: URI::MailTo::EMAIL_REGEXP}
validates :username, uniqueness: {case_sensitive: false}, allow_nil: true,
format: {with: /\A[a-zA-Z0-9_-]+\z/, message: "can only contain letters, numbers, underscores, and hyphens"},
length: {minimum: 2, maximum: 30}
validates :password, length: {minimum: 8}, allow_nil: true
validate :no_reserved_claim_names
# Enum - automatically creates scopes (User.active, User.disabled, etc.)
enum :status, { active: 0, disabled: 1, pending_invitation: 2 }
enum :status, {active: 0, disabled: 1, pending_invitation: 2}
# Scopes
scope :admins, -> { where(admin: true) }
@@ -78,6 +77,14 @@ class User < ApplicationRecord
totp.verify(code, drift_behind: 30, drift_ahead: 30)
end
# Console/debug helper: get current TOTP code
def console_totp
return nil unless totp_enabled?
require "rotp"
ROTP::TOTP.new(totp_secret).now
end
def verify_backup_code(code)
return false unless backup_codes.present?
@@ -115,12 +122,7 @@ class User < ApplicationRecord
cache_key = "backup_code_failed_attempts_#{id}"
attempts = Rails.cache.read(cache_key) || 0
if attempts >= 5 # Allow max 5 failed attempts per hour
true
else
# Don't increment here - increment only on failed attempts
false
end
attempts >= 5
end
# Increment failed attempt counter
@@ -224,7 +226,7 @@ class User < ApplicationRecord
reserved_used = parsed_custom_claims.keys.map(&:to_s) & RESERVED_CLAIMS
if reserved_used.any?
errors.add(:custom_claims, "cannot override reserved OIDC claims: #{reserved_used.join(', ')}")
errors.add(:custom_claims, "cannot override reserved OIDC claims: #{reserved_used.join(", ")}")
end
end

View File

@@ -2,5 +2,5 @@ class UserGroup < ApplicationRecord
belongs_to :user
belongs_to :group
validates :user_id, uniqueness: { scope: :group_id }
validates :user_id, uniqueness: {scope: :group_id}
end

View File

@@ -1,12 +1,15 @@
class WebauthnCredential < ApplicationRecord
belongs_to :user
# Set default authenticator_type if not provided
after_initialize :set_default_authenticator_type, if: :new_record?
# Validations
validates :external_id, presence: true, uniqueness: true
validates :public_key, presence: true
validates :sign_count, presence: true, numericality: { greater_than_or_equal_to: 0, only_integer: true }
validates :sign_count, presence: true, numericality: {greater_than_or_equal_to: 0, only_integer: true}
validates :nickname, presence: true
validates :authenticator_type, inclusion: { in: %w[platform cross-platform] }
validates :authenticator_type, inclusion: {in: %w[platform cross-platform]}
# Scopes for querying
scope :active, -> { where(nil) } # All credentials are active (we can add revoked_at later if needed)
@@ -77,6 +80,10 @@ class WebauthnCredential < ApplicationRecord
private
def set_default_authenticator_type
self.authenticator_type ||= "cross-platform"
end
def time_ago_in_words(time)
seconds = Time.current - time
minutes = seconds / 60
@@ -84,11 +91,11 @@ class WebauthnCredential < ApplicationRecord
days = hours / 24
if days > 0
"#{days.floor} day#{'s' if days > 1} ago"
"#{days.floor} day#{"s" if days > 1} ago"
elsif hours > 0
"#{hours.floor} hour#{'s' if hours > 1} ago"
"#{hours.floor} hour#{"s" if hours > 1} ago"
elsif minutes > 0
"#{minutes.floor} minute#{'s' if minutes > 1} ago"
"#{minutes.floor} minute#{"s" if minutes > 1} ago"
else
"Just now"
end

View File

@@ -13,20 +13,20 @@ module ClaimsMerger
result = base.dup
incoming.each do |key, value|
if result.key?(key)
result[key] = if result.key?(key)
# If both values are arrays, combine them (union to avoid duplicates)
if result[key].is_a?(Array) && value.is_a?(Array)
result[key] = (result[key] + value).uniq
(result[key] + value).uniq
# If both values are hashes, recursively merge them
elsif result[key].is_a?(Hash) && value.is_a?(Hash)
result[key] = deep_merge_claims(result[key], value)
deep_merge_claims(result[key], value)
else
# Otherwise, incoming value wins (override)
result[key] = value
value
end
else
# New key, just add it
result[key] = value
value
end
end

View File

@@ -3,7 +3,7 @@ class OidcJwtService
class << self
# Generate an ID token (JWT) for the user
def generate_id_token(user, application, consent: nil, nonce: nil)
def generate_id_token(user, application, consent: nil, nonce: nil, access_token: nil, auth_time: nil, acr: nil, scopes: "openid")
now = Time.current.to_i
# Use application's configured ID token TTL (defaults to 1 hour)
ttl = application.id_token_expiry_seconds
@@ -11,27 +11,58 @@ class OidcJwtService
# Use pairwise SID from consent if available, fallback to user ID
subject = consent&.sid || user.id.to_s
# Parse scopes (space-separated string)
requested_scopes = scopes.to_s.split
# Required claims (always included per OIDC Core spec)
payload = {
iss: issuer_url,
sub: subject,
aud: application.client_id,
exp: now + ttl,
iat: now,
email: user.email_address,
email_verified: true,
preferred_username: user.username.presence || user.email_address,
name: user.name.presence || user.email_address
iat: now
}
# Email claims (only if 'email' scope requested)
if requested_scopes.include?("email")
payload[:email] = user.email_address
payload[:email_verified] = true
end
# Profile claims (only if 'profile' scope requested)
if requested_scopes.include?("profile")
payload[:preferred_username] = user.username.presence || user.email_address
payload[:name] = user.name.presence || user.email_address
end
# Add nonce if provided (OIDC requires this for implicit flow)
payload[:nonce] = nonce if nonce.present?
# Add groups if user has any
if user.groups.any?
# Add auth_time if provided (OIDC Core §2 - required when max_age is used)
payload[:auth_time] = auth_time if auth_time.present?
# Add acr if provided (OIDC Core §2 - authentication context class reference)
payload[:acr] = acr if acr.present?
# Add azp (authorized party) - the client_id this token was issued to
# OIDC Core §2 - required when aud has multiple values, optional but useful for single
payload[:azp] = application.client_id
# Add at_hash if access token is provided (OIDC Core spec §3.1.3.6)
# at_hash = left-most 128 bits of SHA-256 hash of access token, base64url encoded
if access_token.present?
sha256 = Digest::SHA256.digest(access_token)
at_hash = Base64.urlsafe_encode64(sha256[0..15], padding: false)
payload[:at_hash] = at_hash
end
# Groups claims (only if 'groups' scope requested)
if requested_scopes.include?("groups") && user.groups.any?
payload[:groups] = user.groups.pluck(:name)
end
# Merge custom claims from groups (arrays are combined, not overwritten)
# Note: Custom claims from groups are always merged (not scope-dependent)
user.groups.each do |group|
payload = deep_merge_claims(payload, group.parsed_custom_claims)
end
@@ -42,12 +73,36 @@ class OidcJwtService
# Merge app-specific custom claims (highest priority, arrays are combined)
payload = deep_merge_claims(payload, application.custom_claims_for_user(user))
JWT.encode(payload, private_key, "RS256", { kid: key_id, typ: "JWT" })
JWT.encode(payload, private_key, "RS256", {kid: key_id, typ: "JWT"})
end
# Generate a backchannel logout token (JWT)
# Per OIDC Back-Channel Logout spec, this token:
# - MUST include iss, aud, iat, jti, events claims
# - MUST include sub or sid (or both) - we always include both
# - MUST NOT include nonce claim
def generate_logout_token(user, application, consent)
now = Time.current.to_i
payload = {
iss: issuer_url,
sub: consent.sid, # Pairwise subject identifier
aud: application.client_id,
iat: now,
jti: SecureRandom.uuid, # Unique identifier for this logout token
sid: consent.sid, # Session ID - always included for granular logout
events: {
"http://schemas.openid.net/event/backchannel-logout" => {}
}
}
# Important: Do NOT include nonce in logout tokens (spec requirement)
JWT.encode(payload, private_key, "RS256", {kid: key_id, typ: "JWT"})
end
# Decode and verify an ID token
def decode_id_token(token)
JWT.decode(token, public_key, true, { algorithm: "RS256" })
JWT.decode(token, public_key, true, {algorithm: "RS256"})
end
# Get the public key in JWK format for the JWKS endpoint

View File

@@ -17,6 +17,87 @@
<%= form.text_area :description, rows: 3, class: "mt-1 block w-full rounded-md border-gray-300 shadow-sm focus:border-blue-500 focus:ring-blue-500 sm:text-sm", placeholder: "Optional description of this application" %>
</div>
<div>
<div class="flex items-center justify-between">
<%= form.label :icon, "Application Icon", class: "block text-sm font-medium text-gray-700" %>
<a href="https://dashboardicons.com" target="_blank" rel="noopener noreferrer" class="text-xs text-blue-600 hover:text-blue-800 flex items-center gap-1">
<svg class="w-3 h-3" fill="none" stroke="currentColor" viewBox="0 0 24 24">
<path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M10 6H6a2 2 0 00-2 2v10a2 2 0 002 2h10a2 2 0 002-2v-4M14 4h6m0 0v6m0-6L10 14"></path>
</svg>
Browse icons at dashboardicons.com
</a>
</div>
<% if application.icon.attached? && application.persisted? %>
<% begin %>
<%# Only show icon if we can successfully get its URL (blob is persisted) %>
<% if application.icon.blob&.persisted? && application.icon.blob.key.present? %>
<div class="mt-2 mb-3 flex items-center gap-4">
<%= image_tag application.icon, class: "h-16 w-16 rounded-lg object-cover border border-gray-200", alt: "Current icon" %>
<div class="text-sm text-gray-600">
<p class="font-medium">Current icon</p>
<p class="text-xs"><%= number_to_human_size(application.icon.blob.byte_size) %></p>
</div>
</div>
<% end %>
<% rescue ArgumentError => e %>
<%# Handle case where icon attachment exists but can't generate signed_id %>
<% if e.message.include?("Cannot get a signed_id for a new record") %>
<div class="mt-2 mb-3 text-sm text-gray-600">
<p class="font-medium">Icon uploaded</p>
<p class="text-xs">File will be processed shortly</p>
</div>
<% else %>
<%# Re-raise if it's a different error %>
<% raise e %>
<% end %>
<% end %>
<% end %>
<div class="mt-2" data-controller="file-drop image-paste">
<div class="flex justify-center px-6 pt-5 pb-6 border-2 border-gray-300 border-dashed rounded-md hover:border-blue-400 transition-colors"
data-file-drop-target="dropzone"
data-image-paste-target="dropzone"
data-action="dragover->file-drop#dragover dragleave->file-drop#dragleave drop->file-drop#drop paste->image-paste#handlePaste"
tabindex="0">
<div class="space-y-1 text-center">
<svg class="mx-auto h-12 w-12 text-gray-400" stroke="currentColor" fill="none" viewBox="0 0 48 48">
<path d="M28 8H12a4 4 0 00-4 4v20m32-12v8m0 0v8a4 4 0 01-4 4H12a4 4 0 01-4-4v-4m32-4l-3.172-3.172a4 4 0 00-5.656 0L28 28M8 32l9.172-9.172a4 4 0 015.656 0L28 28m0 0l4 4m4-24h8m-4-4v8m-12 4h.02" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" />
</svg>
<div class="flex text-sm text-gray-600">
<label for="<%= form.field_id(:icon) %>" class="relative cursor-pointer bg-white rounded-md font-medium text-blue-600 hover:text-blue-500 focus-within:outline-none focus-within:ring-2 focus-within:ring-offset-2 focus-within:ring-blue-500">
<span>Upload a file</span>
<%= form.file_field :icon,
accept: "image/png,image/jpg,image/jpeg,image/gif,image/svg+xml",
class: "sr-only",
data: {
file_drop_target: "input",
image_paste_target: "input",
action: "change->file-drop#handleFiles"
} %>
</label>
<p class="pl-1">or drag and drop</p>
</div>
<p class="text-xs text-gray-500">PNG, JPG, GIF, or SVG up to 2MB</p>
<p class="text-xs text-blue-600 font-medium mt-2">💡 Tip: Click here and press Ctrl+V (or Cmd+V) to paste an image from your clipboard</p>
</div>
</div>
<div data-file-drop-target="preview" class="mt-3 hidden">
<div class="flex items-center gap-3 p-3 bg-blue-50 rounded-md border border-blue-200">
<img data-file-drop-target="previewImage" class="h-12 w-12 rounded object-cover" alt="Preview">
<div class="flex-1 min-w-0">
<p class="text-sm font-medium text-gray-900" data-file-drop-target="filename"></p>
<p class="text-xs text-gray-500" data-file-drop-target="filesize"></p>
</div>
<button type="button" data-action="click->file-drop#clear" class="text-gray-400 hover:text-gray-600">
<svg class="h-5 w-5" fill="currentColor" viewBox="0 0 20 20">
<path fill-rule="evenodd" d="M4.293 4.293a1 1 0 011.414 0L10 8.586l4.293-4.293a1 1 0 111.414 1.414L11.414 10l4.293 4.293a1 1 0 01-1.414 1.414L10 11.414l-4.293 4.293a1 1 0 01-1.414-1.414L8.586 10 4.293 5.707a1 1 0 010-1.414z" clip-rule="evenodd" />
</svg>
</button>
</div>
</div>
</div>
</div>
<div>
<%= form.label :landing_url, "Landing URL", class: "block text-sm font-medium text-gray-700" %>
<%= form.url_field :landing_url, class: "mt-1 block w-full rounded-md border-gray-300 shadow-sm focus:border-blue-500 focus:ring-blue-500 sm:text-sm", placeholder: "https://app.example.com" %>
@@ -39,12 +120,67 @@
<div id="oidc-fields" class="space-y-6 border-t border-gray-200 pt-6 <%= 'hidden' unless application.oidc? || !application.persisted? %>" data-application-form-target="oidcFields">
<h3 class="text-base font-semibold text-gray-900">OIDC Configuration</h3>
<!-- Client Type Selection (only for new applications) -->
<% unless application.persisted? %>
<div class="border border-gray-200 rounded-lg p-4 bg-gray-50">
<h4 class="text-sm font-semibold text-gray-900 mb-3">Client Type</h4>
<div class="space-y-3">
<div class="flex items-start">
<%= form.radio_button :is_public_client, "false", checked: !application.is_public_client, class: "mt-1 h-4 w-4 border-gray-300 text-blue-600 focus:ring-blue-500", data: { action: "change->application-form#updatePkceVisibility" } %>
<div class="ml-3">
<label for="application_is_public_client_false" class="block text-sm font-medium text-gray-900">Confidential Client (Recommended)</label>
<p class="text-sm text-gray-500">Backend server app that can securely store a client secret. Examples: traditional web apps, server-to-server APIs.</p>
</div>
</div>
<div class="flex items-start">
<%= form.radio_button :is_public_client, "true", checked: application.is_public_client, class: "mt-1 h-4 w-4 border-gray-300 text-blue-600 focus:ring-blue-500", data: { action: "change->application-form#updatePkceVisibility" } %>
<div class="ml-3">
<label for="application_is_public_client_true" class="block text-sm font-medium text-gray-900">Public Client</label>
<p class="text-sm text-gray-500">Frontend-only app that cannot store secrets securely. Examples: SPAs (React/Vue), mobile apps, CLI tools. <strong class="text-amber-600">PKCE is required.</strong></p>
</div>
</div>
</div>
</div>
<% else %>
<!-- Show client type for existing applications (read-only) -->
<div class="flex items-center gap-2 text-sm">
<span class="font-medium text-gray-700">Client Type:</span>
<% if application.public_client? %>
<span class="inline-flex items-center rounded-md bg-amber-50 px-2 py-1 text-xs font-medium text-amber-700 ring-1 ring-inset ring-amber-600/20">Public Client (PKCE Required)</span>
<% else %>
<span class="inline-flex items-center rounded-md bg-green-50 px-2 py-1 text-xs font-medium text-green-700 ring-1 ring-inset ring-green-600/20">Confidential Client</span>
<% end %>
</div>
<% end %>
<!-- PKCE Requirement (only for confidential clients) -->
<div id="pkce-options" data-application-form-target="pkceOptions" class="<%= 'hidden' if application.persisted? && application.public_client? %>">
<div class="flex items-center">
<%= form.check_box :require_pkce, class: "h-4 w-4 rounded border-gray-300 text-blue-600 focus:ring-blue-500" %>
<%= form.label :require_pkce, "Require PKCE (Proof Key for Code Exchange)", class: "ml-2 block text-sm font-medium text-gray-900" %>
</div>
<p class="ml-6 text-sm text-gray-500">
Recommended for enhanced security (OAuth 2.1 best practice).
<br><span class="text-xs text-gray-400">Note: Public clients always require PKCE regardless of this setting.</span>
</p>
</div>
<div>
<%= form.label :redirect_uris, "Redirect URIs", class: "block text-sm font-medium text-gray-700" %>
<%= form.text_area :redirect_uris, rows: 4, class: "mt-1 block w-full rounded-md border-gray-300 shadow-sm focus:border-blue-500 focus:ring-blue-500 sm:text-sm font-mono", placeholder: "https://example.com/callback\nhttps://app.example.com/auth/callback" %>
<p class="mt-1 text-sm text-gray-500">One URI per line. These are the allowed callback URLs for your application.</p>
</div>
<div>
<%= form.label :backchannel_logout_uri, "Backchannel Logout URI (Optional)", class: "block text-sm font-medium text-gray-700" %>
<%= form.url_field :backchannel_logout_uri, class: "mt-1 block w-full rounded-md border-gray-300 shadow-sm focus:border-blue-500 focus:ring-blue-500 sm:text-sm font-mono", placeholder: "https://app.example.com/oidc/backchannel-logout" %>
<p class="mt-1 text-sm text-gray-500">
If the application supports OpenID Connect Backchannel Logout, enter the logout endpoint URL.
When users log out, Clinch will send logout notifications to this endpoint for immediate session termination.
Leave blank if the application doesn't support backchannel logout.
</p>
</div>
<div class="border-t border-gray-200 pt-4 mt-4">
<h4 class="text-sm font-semibold text-gray-900 mb-3">Token Expiration Settings</h4>
<p class="text-sm text-gray-500 mb-4">Configure how long tokens remain valid. Shorter times are more secure but require more frequent refreshes.</p>

View File

@@ -14,7 +14,7 @@
<table class="min-w-full divide-y divide-gray-300">
<thead>
<tr>
<th scope="col" class="py-3.5 pl-4 pr-3 text-left text-sm font-semibold text-gray-900 sm:pl-0">Name</th>
<th scope="col" class="py-3.5 pl-4 pr-3 text-left text-sm font-semibold text-gray-900 sm:pl-0">Application</th>
<th scope="col" class="px-3 py-3.5 text-left text-sm font-semibold text-gray-900">Slug</th>
<th scope="col" class="px-3 py-3.5 text-left text-sm font-semibold text-gray-900">Type</th>
<th scope="col" class="px-3 py-3.5 text-left text-sm font-semibold text-gray-900">Status</th>
@@ -28,7 +28,18 @@
<% @applications.each do |application| %>
<tr>
<td class="whitespace-nowrap py-4 pl-4 pr-3 text-sm font-medium text-gray-900 sm:pl-0">
<div class="flex items-center gap-3">
<% if application.icon.attached? %>
<%= image_tag application.icon, class: "h-10 w-10 rounded-lg object-cover border border-gray-200 flex-shrink-0", alt: "#{application.name} icon" %>
<% else %>
<div class="h-10 w-10 rounded-lg bg-gray-100 border border-gray-200 flex items-center justify-center flex-shrink-0">
<svg class="h-6 w-6 text-gray-400" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M4 16l4.586-4.586a2 2 0 012.828 0L16 16m-2-2l1.586-1.586a2 2 0 012.828 0L20 14m-6-6h.01M6 20h12a2 2 0 002-2V6a2 2 0 00-2-2H6a2 2 0 00-2 2v12a2 2 0 002 2z" />
</svg>
</div>
<% end %>
<%= link_to application.name, admin_application_path(application), class: "text-blue-600 hover:text-blue-900" %>
</div>
</td>
<td class="whitespace-nowrap px-3 py-4 text-sm text-gray-500">
<code class="text-xs bg-gray-100 px-2 py-1 rounded"><%= application.slug %></code>

View File

@@ -1,26 +1,50 @@
<div class="mb-6">
<% if flash[:client_id] && flash[:client_secret] %>
<% if flash[:client_id] %>
<div class="bg-yellow-50 border border-yellow-200 rounded-md p-4 mb-6">
<h4 class="text-sm font-medium text-yellow-800 mb-2">🔐 OIDC Client Credentials</h4>
<% if flash[:public_client] %>
<p class="text-xs text-yellow-700 mb-3">This is a public client. Copy the client ID below.</p>
<% else %>
<p class="text-xs text-yellow-700 mb-3">Copy these credentials now. The client secret will not be shown again.</p>
<% end %>
<div class="space-y-2">
<div>
<span class="text-xs font-medium text-yellow-700">Client ID:</span>
</div>
<code class="block bg-yellow-100 px-3 py-2 rounded font-mono text-xs break-all"><%= flash[:client_id] %></code>
<% if flash[:client_secret] %>
<div class="mt-3">
<span class="text-xs font-medium text-yellow-700">Client Secret:</span>
</div>
<code class="block bg-yellow-100 px-3 py-2 rounded font-mono text-xs break-all"><%= flash[:client_secret] %></code>
<% elsif flash[:public_client] %>
<div class="mt-3">
<span class="text-xs font-medium text-yellow-700">Client Secret:</span>
</div>
<div class="bg-yellow-100 px-3 py-2 rounded text-xs text-yellow-600">
Public clients do not have a client secret. PKCE is required.
</div>
<% end %>
</div>
</div>
<% end %>
<div class="sm:flex sm:items-center sm:justify-between">
<div class="sm:flex sm:items-start sm:justify-between">
<div class="flex items-start gap-4">
<% if @application.icon.attached? %>
<%= image_tag @application.icon, class: "h-16 w-16 rounded-lg object-cover border border-gray-200 shrink-0", alt: "#{@application.name} icon" %>
<% else %>
<div class="h-16 w-16 rounded-lg bg-gray-100 border border-gray-200 flex items-center justify-center shrink-0">
<svg class="h-8 w-8 text-gray-400" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M4 16l4.586-4.586a2 2 0 012.828 0L16 16m-2-2l1.586-1.586a2 2 0 012.828 0L20 14m-6-6h.01M6 20h12a2 2 0 002-2V6a2 2 0 00-2-2H6a2 2 0 00-2 2v12a2 2 0 002 2z" />
</svg>
</div>
<% end %>
<div>
<h1 class="text-2xl font-semibold text-gray-900"><%= @application.name %></h1>
<p class="mt-1 text-sm text-gray-500"><%= @application.description %></p>
</div>
</div>
<div class="mt-4 sm:mt-0 flex gap-3">
<%= link_to "Edit", edit_admin_application_path(@application), class: "rounded-md bg-white px-3 py-2 text-sm font-semibold text-gray-900 shadow-sm ring-1 ring-inset ring-gray-300 hover:bg-gray-50" %>
<%= button_to "Delete", admin_application_path(@application), method: :delete, data: { turbo_confirm: "Are you sure?" }, class: "rounded-md bg-red-600 px-3 py-2 text-sm font-semibold text-white shadow-sm hover:bg-red-500" %>
@@ -78,16 +102,40 @@
<div class="bg-white shadow sm:rounded-lg">
<div class="px-4 py-5 sm:p-6">
<div class="flex items-center justify-between mb-4">
<h3 class="text-base font-semibold leading-6 text-gray-900">OIDC Credentials</h3>
<h3 class="text-base font-semibold leading-6 text-gray-900">OIDC Configuration</h3>
<%= button_to "Regenerate Credentials", regenerate_credentials_admin_application_path(@application), method: :post, data: { turbo_confirm: "This will invalidate the current credentials. Continue?" }, class: "text-sm text-red-600 hover:text-red-900" %>
</div>
<dl class="space-y-4">
<div class="grid grid-cols-2 gap-4">
<div>
<dt class="text-sm font-medium text-gray-500">Client Type</dt>
<dd class="mt-1 text-sm text-gray-900">
<% if @application.public_client? %>
<span class="inline-flex items-center rounded-full bg-blue-100 px-2 py-1 text-xs font-medium text-blue-700">Public</span>
<% else %>
<span class="inline-flex items-center rounded-full bg-gray-100 px-2 py-1 text-xs font-medium text-gray-700">Confidential</span>
<% end %>
</dd>
</div>
<div>
<dt class="text-sm font-medium text-gray-500">PKCE</dt>
<dd class="mt-1 text-sm text-gray-900">
<% if @application.requires_pkce? %>
<span class="inline-flex items-center rounded-full bg-green-100 px-2 py-1 text-xs font-medium text-green-700">Required</span>
<% else %>
<span class="inline-flex items-center rounded-full bg-gray-100 px-2 py-1 text-xs font-medium text-gray-700">Optional</span>
<% end %>
</dd>
</div>
</div>
<% unless flash[:client_id] %>
<div>
<dt class="text-sm font-medium text-gray-500">Client ID</dt>
<dd class="mt-1 text-sm text-gray-900">
<code class="block bg-gray-100 px-3 py-2 rounded font-mono text-xs break-all"><%= @application.client_id %></code>
</dd>
</div>
<% if @application.confidential_client? %>
<div>
<dt class="text-sm font-medium text-gray-500">Client Secret</dt>
<dd class="mt-1 text-sm text-gray-900">
@@ -99,6 +147,17 @@
</p>
</dd>
</div>
<% else %>
<div>
<dt class="text-sm font-medium text-gray-500">Client Secret</dt>
<dd class="mt-1 text-sm text-gray-900">
<div class="bg-blue-50 px-3 py-2 rounded text-xs text-blue-600">
Public clients do not use a client secret. PKCE is required for authorization.
</div>
</dd>
</div>
<% end %>
<% end %>
<div>
<dt class="text-sm font-medium text-gray-500">Redirect URIs</dt>
<dd class="mt-1 text-sm text-gray-900">
@@ -111,6 +170,27 @@
<% end %>
</dd>
</div>
<div>
<dt class="text-sm font-medium text-gray-500">
Backchannel Logout URI
<% if @application.supports_backchannel_logout? %>
<span class="ml-2 inline-flex items-center rounded-full bg-green-100 px-2 py-0.5 text-xs font-medium text-green-700">Enabled</span>
<% end %>
</dt>
<dd class="mt-1 text-sm text-gray-900">
<% if @application.backchannel_logout_uri.present? %>
<code class="block bg-gray-100 px-3 py-2 rounded font-mono text-xs break-all"><%= @application.backchannel_logout_uri %></code>
<p class="mt-2 text-xs text-gray-500">
When users log out, Clinch will send logout notifications to this endpoint for immediate session termination.
</p>
<% else %>
<span class="text-gray-400 italic">Not configured</span>
<p class="mt-1 text-xs text-gray-500">
Backchannel logout is optional. Configure it if the application supports OpenID Connect Backchannel Logout.
</p>
<% end %>
</dd>
</div>
</dl>
</div>
</div>

View File

@@ -102,11 +102,22 @@
<% @applications.each do |app| %>
<div class="bg-white rounded-lg border border-gray-200 shadow-sm hover:shadow-md transition">
<div class="p-6">
<div class="flex items-center justify-between mb-3">
<div class="flex items-start gap-3 mb-4">
<% if app.icon.attached? %>
<%= image_tag app.icon, class: "h-12 w-12 rounded-lg object-cover border border-gray-200 shrink-0", alt: "#{app.name} icon" %>
<% else %>
<div class="h-12 w-12 rounded-lg bg-gray-100 border border-gray-200 flex items-center justify-center shrink-0">
<svg class="h-6 w-6 text-gray-400" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M4 16l4.586-4.586a2 2 0 012.828 0L16 16m-2-2l1.586-1.586a2 2 0 012.828 0L20 14m-6-6h.01M6 20h12a2 2 0 002-2V6a2 2 0 00-2-2H6a2 2 0 00-2 2v12a2 2 0 002 2z" />
</svg>
</div>
<% end %>
<div class="flex-1 min-w-0">
<div class="flex items-start justify-between">
<h3 class="text-lg font-semibold text-gray-900 truncate">
<%= app.name %>
</h3>
<span class="inline-flex items-center px-2.5 py-0.5 rounded-full text-xs font-medium
<span class="ml-2 inline-flex items-center px-2.5 py-0.5 rounded-full text-xs font-medium shrink-0
<% if app.oidc? %>
bg-blue-100 text-blue-800
<% else %>
@@ -115,15 +126,15 @@
<%= app.app_type.humanize %>
</span>
</div>
<p class="text-sm text-gray-600 mb-4">
<% if app.oidc? %>
OIDC Application
<% else %>
ForwardAuth Protected Application
<% end %>
<% if app.description.present? %>
<p class="text-sm text-gray-600 mt-1 line-clamp-2">
<%= app.description %>
</p>
<% end %>
</div>
</div>
<div class="space-y-2">
<% if app.landing_url.present? %>
<%= link_to "Open Application", app.landing_url,
target: "_blank",
@@ -134,6 +145,13 @@
No landing URL configured
</div>
<% end %>
<% if app.user_has_active_session?(@user) %>
<%= button_to "Logout", logout_from_app_active_sessions_path(application_id: app.id), method: :delete,
class: "w-full flex justify-center items-center px-4 py-2 border border-orange-300 text-sm font-medium rounded-md text-orange-700 bg-white hover:bg-orange-50 focus:outline-none focus:ring-2 focus:ring-offset-2 focus:ring-orange-500 transition",
form: { data: { turbo_confirm: "This will log you out of #{app.name}. You can sign back in without re-authorizing. Continue?" } } %>
<% end %>
</div>
</div>
</div>
<% end %>

View File

@@ -1,6 +1,15 @@
<div class="mx-auto max-w-md">
<div class="bg-white py-8 px-6 shadow rounded-lg sm:px-10">
<div class="mb-8">
<div class="mb-8 text-center">
<% if @application.icon.attached? %>
<%= image_tag @application.icon, class: "mx-auto h-20 w-20 rounded-xl object-cover border-2 border-gray-200 shadow-sm mb-4", alt: "#{@application.name} icon" %>
<% else %>
<div class="mx-auto h-20 w-20 rounded-xl bg-gray-100 border-2 border-gray-200 flex items-center justify-center mb-4">
<svg class="h-10 w-10 text-gray-400" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M4 16l4.586-4.586a2 2 0 012.828 0L16 16m-2-2l1.586-1.586a2 2 0 012.828 0L20 14m-6-6h.01M6 20h12a2 2 0 002-2V6a2 2 0 00-2-2H6a2 2 0 00-2 2v12a2 2 0 002 2z" />
</svg>
</div>
<% end %>
<h2 class="text-2xl font-bold text-gray-900">Authorize Application</h2>
<p class="mt-2 text-sm text-gray-600">
<strong><%= @application.name %></strong> is requesting access to your account.

View File

@@ -31,6 +31,15 @@
class: "mt-1 block w-full rounded-md border-gray-300 shadow-sm focus:border-blue-500 focus:ring-blue-500 sm:text-sm" %>
</div>
<div>
<%= form.label :current_password, "Current Password", class: "block text-sm font-medium text-gray-700" %>
<%= form.password_field :current_password,
autocomplete: "current-password",
placeholder: "Required to change email",
class: "mt-1 block w-full rounded-md border-gray-300 shadow-sm focus:border-blue-500 focus:ring-blue-500 sm:text-sm" %>
<p class="mt-1 text-sm text-gray-500">Enter your current password to confirm this change</p>
</div>
<div>
<%= form.submit "Update Email", class: "inline-flex justify-center rounded-md border border-transparent bg-blue-600 py-2 px-4 text-sm font-medium text-white shadow-sm hover:bg-blue-700 focus:outline-none focus:ring-2 focus:ring-blue-500 focus:ring-offset-2" %>
</div>

View File

@@ -1,6 +1,8 @@
<%# Enhanced Flash Messages with Support for Multiple Types and Auto-Dismiss %>
<% flash.each do |type, message| %>
<% next if message.blank? %>
<%# Skip credential-related flash messages - they're displayed in a special credentials box %>
<% next if %w[client_id client_secret public_client].include?(type.to_s) %>
<%
# Map flash types to styling

View File

@@ -2,6 +2,4 @@
require "rubygems"
require "bundler/setup"
ARGV.unshift("--ensure-latest")
load Gem.bin_path("brakeman", "brakeman")

5
bin/standardrb Executable file
View File

@@ -0,0 +1,5 @@
#!/usr/bin/env ruby
require "rubygems"
require "bundler/setup"
load Gem.bin_path("standard", "standardrb")

View File

@@ -27,13 +27,13 @@ module Clinch
# Configure SMTP settings using environment variables
config.action_mailer.delivery_method = :smtp
config.action_mailer.smtp_settings = {
address: ENV.fetch('SMTP_ADDRESS', 'localhost'),
port: ENV.fetch('SMTP_PORT', 587),
domain: ENV.fetch('SMTP_DOMAIN', 'localhost'),
user_name: ENV.fetch('SMTP_USERNAME', nil),
password: ENV.fetch('SMTP_PASSWORD', nil),
authentication: ENV.fetch('SMTP_AUTHENTICATION', 'plain').to_sym,
enable_starttls_auto: ENV.fetch('SMTP_STARTTLS_AUTO', 'true') == 'true',
address: ENV.fetch("SMTP_ADDRESS", "localhost"),
port: ENV.fetch("SMTP_PORT", 587),
domain: ENV.fetch("SMTP_DOMAIN", "localhost"),
user_name: ENV.fetch("SMTP_USERNAME", nil),
password: ENV.fetch("SMTP_PASSWORD", nil),
authentication: ENV.fetch("SMTP_AUTHENTICATION", "plain").to_sym,
enable_starttls_auto: ENV.fetch("SMTP_STARTTLS_AUTO", "true") == "true",
openssl_verify_mode: OpenSSL::SSL::VERIFY_PEER
}
end

View File

@@ -20,7 +20,7 @@ Rails.application.configure do
if Rails.root.join("tmp/caching-dev.txt").exist?
config.action_controller.perform_caching = true
config.action_controller.enable_fragment_cache_logging = true
config.public_file_server.headers = { "cache-control" => "public, max-age=#{2.days.to_i}" }
config.public_file_server.headers = {"cache-control" => "public, max-age=#{2.days.to_i}"}
else
config.action_controller.perform_caching = false
end
@@ -39,10 +39,10 @@ Rails.application.configure do
config.action_mailer.perform_caching = false
# Set localhost to be used by links generated in mailer templates.
config.action_mailer.default_url_options = { host: "localhost", port: 3000 }
config.action_mailer.default_url_options = {host: "localhost", port: 3000}
# Log with request_id as a tag (same as production).
config.log_tags = [ :request_id ]
config.log_tags = [:request_id]
# Print deprecation notices to the Rails logger.
config.active_support.deprecation = :log
@@ -62,7 +62,6 @@ Rails.application.configure do
# Use async processor for background jobs in development
config.active_job.queue_adapter = :async
# Highlight code that triggered redirect in logs.
config.action_dispatch.verbose_redirect_logs = true

View File

@@ -16,7 +16,7 @@ Rails.application.configure do
config.action_controller.perform_caching = true
# Cache assets for far-future expiry since they are all digest stamped.
config.public_file_server.headers = { "cache-control" => "public, max-age=#{1.year.to_i}" }
config.public_file_server.headers = {"cache-control" => "public, max-age=#{1.year.to_i}"}
# Enable serving of images, stylesheets, and JavaScripts from an asset server.
# config.asset_host = "http://assets.example.com"
@@ -30,12 +30,20 @@ Rails.application.configure do
# Force all access to the app over SSL, use Strict-Transport-Security, and use secure cookies.
config.force_ssl = true
# Additional security headers (beyond Rails defaults)
# Note: Rails already sets X-Content-Type-Options: nosniff by default
# Note: Permissions-Policy is configured in config/initializers/permissions_policy.rb
config.action_dispatch.default_headers.merge!(
"X-Frame-Options" => "DENY", # Override default SAMEORIGIN to prevent clickjacking
"Referrer-Policy" => "strict-origin-when-cross-origin" # Control referrer information
)
# Skip http-to-https redirect for the default health check endpoint.
# config.ssl_options = { redirect: { exclude: ->(request) { request.path == "/up" } } }
# Log to STDOUT with the current request id as a default log tag.
config.log_tags = [ :request_id ]
config.logger = ActiveSupport::TaggedLogging.logger(STDOUT)
config.log_tags = [:request_id]
config.logger = ActiveSupport::TaggedLogging.logger($stdout)
# Change to "debug" to log everything (including potentially personally-identifiable information!).
config.log_level = ENV.fetch("RAILS_LOG_LEVEL", "info")
@@ -49,8 +57,8 @@ Rails.application.configure do
# Replace the default in-process memory cache store with a durable alternative.
config.cache_store = :solid_cache_store
# Use async processor for background jobs (modify as needed for production)
config.active_job.queue_adapter = :async
# Use Solid Queue for background jobs
config.active_job.queue_adapter = :solid_queue
# Ignore bad email addresses and do not raise email delivery errors.
# Set this to true and configure the email server for immediate delivery to raise delivery errors.
@@ -58,7 +66,7 @@ Rails.application.configure do
# Set host to be used by links generated in mailer templates.
config.action_mailer.default_url_options = {
host: ENV.fetch('CLINCH_HOST', 'example.com')
host: ENV.fetch("CLINCH_HOST", "example.com")
}
# Specify outgoing SMTP server. Remember to add smtp/* credentials via bin/rails credentials:edit.
@@ -78,13 +86,13 @@ Rails.application.configure do
config.active_record.dump_schema_after_migration = false
# Only use :id for inspections in production.
config.active_record.attributes_for_inspect = [ :id ]
config.active_record.attributes_for_inspect = [:id]
# Helper method to extract domain from CLINCH_HOST (removes protocol if present)
def self.extract_domain(host)
return host if host.blank?
# Remove protocol (http:// or https://) if present
host.gsub(/^https?:\/\//, '')
host.gsub(/^https?:\/\//, "")
end
# Helper method to ensure URL has https:// protocol
@@ -97,11 +105,11 @@ Rails.application.configure do
# Enable DNS rebinding protection and other `Host` header attacks.
# Configure allowed hosts based on deployment scenario
allowed_hosts = [
extract_domain(ENV.fetch('CLINCH_HOST', 'auth.example.com')), # External domain (auth service itself)
extract_domain(ENV.fetch("CLINCH_HOST", "auth.example.com")) # External domain (auth service itself)
]
# Use PublicSuffix to extract registrable domain and allow all subdomains
host_domain = extract_domain(ENV.fetch('CLINCH_HOST', 'auth.example.com'))
host_domain = extract_domain(ENV.fetch("CLINCH_HOST", "auth.example.com"))
if host_domain.present?
begin
# Use PublicSuffix to properly extract the domain
@@ -115,20 +123,20 @@ Rails.application.configure do
rescue PublicSuffix::DomainInvalid
# Fallback to simple domain extraction if PublicSuffix fails
Rails.logger.warn "Could not parse domain '#{host_domain}' with PublicSuffix, using fallback"
base_domain = host_domain.split('.').last(2).join('.')
base_domain = host_domain.split(".").last(2).join(".")
allowed_hosts << /.*#{Regexp.escape(base_domain)}/
end
end
# Allow Docker service names if running in same compose
if ENV['CLINCH_DOCKER_SERVICE_NAME']
allowed_hosts << ENV['CLINCH_DOCKER_SERVICE_NAME']
if ENV["CLINCH_DOCKER_SERVICE_NAME"]
allowed_hosts << ENV["CLINCH_DOCKER_SERVICE_NAME"]
end
# Allow internal IP access for cross-compose or host networking
if ENV['CLINCH_ALLOW_INTERNAL_IPS'] == 'true'
if ENV["CLINCH_ALLOW_INTERNAL_IPS"] == "true"
# Specific host IP
allowed_hosts << '192.168.2.246'
allowed_hosts << "192.168.2.246"
# Private IP ranges for internal network access
allowed_hosts += [
@@ -139,14 +147,14 @@ Rails.application.configure do
end
# Local development fallbacks
if ENV['CLINCH_ALLOW_LOCALHOST'] == 'true'
allowed_hosts += ['localhost', '127.0.0.1', '0.0.0.0']
if ENV["CLINCH_ALLOW_LOCALHOST"] == "true"
allowed_hosts += ["localhost", "127.0.0.1", "0.0.0.0"]
end
config.hosts = allowed_hosts
# Skip DNS rebinding protection for the default health check endpoint.
config.host_authorization = { exclude: ->(request) { request.path == "/up" } }
config.host_authorization = {exclude: ->(request) { request.path == "/up" }}
# Sentry configuration for production
# Only enabled if SENTRY_DSN environment variable is set

View File

@@ -16,7 +16,7 @@ Rails.application.configure do
config.eager_load = ENV["CI"].present?
# Configure public file server for tests with cache-control for performance.
config.public_file_server.headers = { "cache-control" => "public, max-age=3600" }
config.public_file_server.headers = {"cache-control" => "public, max-age=3600"}
# Show full error reports.
config.consider_all_requests_local = true
@@ -37,7 +37,7 @@ Rails.application.configure do
config.action_mailer.delivery_method = :test
# Set host to be used by links generated in mailer templates.
config.action_mailer.default_url_options = { host: "example.com" }
config.action_mailer.default_url_options = {host: "example.com"}
# Print deprecation notices to the stderr.
config.active_support.deprecation = :stderr

View File

@@ -0,0 +1,28 @@
# ActiveRecord Encryption Configuration
# Encryption keys derived from SECRET_KEY_BASE (no separate key storage needed)
# Used for encrypting sensitive columns (currently: TOTP secrets)
#
# Optional: Override with env vars (for key rotation or explicit key management):
# - ACTIVE_RECORD_ENCRYPTION_PRIMARY_KEY
# - ACTIVE_RECORD_ENCRYPTION_DETERMINISTIC_KEY
# - ACTIVE_RECORD_ENCRYPTION_KEY_DERIVATION_SALT
# Use env vars if set, otherwise derive from SECRET_KEY_BASE (deterministic)
primary_key = ENV.fetch("ACTIVE_RECORD_ENCRYPTION_PRIMARY_KEY") do
Rails.application.key_generator.generate_key("active_record_encryption_primary", 32)
end
deterministic_key = ENV.fetch("ACTIVE_RECORD_ENCRYPTION_DETERMINISTIC_KEY") do
Rails.application.key_generator.generate_key("active_record_encryption_deterministic", 32)
end
key_derivation_salt = ENV.fetch("ACTIVE_RECORD_ENCRYPTION_KEY_DERIVATION_SALT") do
Rails.application.key_generator.generate_key("active_record_encryption_salt", 32)
end
# Configure Rails 7.1+ ActiveRecord encryption
Rails.application.config.active_record.encryption.primary_key = primary_key
Rails.application.config.active_record.encryption.deterministic_key = deterministic_key
Rails.application.config.active_record.encryption.key_derivation_salt = key_derivation_salt
# Allow unencrypted data for existing records (new/updated records will be encrypted)
# Set to false after all existing encrypted columns have been migrated
Rails.application.config.active_record.encryption.support_unencrypted_data = true

View File

@@ -59,7 +59,6 @@ Rails.application.configure do
policy.report_uri "/api/csp-violation-report"
end
# Start with CSP in report-only mode for testing
# Set to false after verifying everything works in production
config.content_security_policy_report_only = Rails.env.development?

View File

@@ -8,7 +8,7 @@ Rails.application.config.after_initialize do
# Configure log rotation
csp_logger = Logger.new(
csp_log_path,
'daily', # Rotate daily
"daily", # Rotate daily
30 # Keep 30 old log files
)
@@ -16,7 +16,7 @@ Rails.application.config.after_initialize do
# Format: [TIMESTAMP] LEVEL MESSAGE
csp_logger.formatter = proc do |severity, datetime, progname, msg|
"[#{datetime.strftime('%Y-%m-%d %H:%M:%S')}] #{severity} #{msg}\n"
"[#{datetime.strftime("%Y-%m-%d %H:%M:%S")}] #{severity} #{msg}\n"
end
module CspViolationLocalLogger
@@ -69,7 +69,6 @@ Rails.application.config.after_initialize do
# Also log to main Rails logger for visibility
Rails.logger.info "CSP violation logged to csp_violations.log: #{violated_directive} - #{blocked_uri}"
rescue => e
# Ensure logger errors don't break the CSP reporting flow
Rails.logger.error "Failed to log CSP violation to file: #{e.message}"
@@ -81,12 +80,12 @@ Rails.application.config.after_initialize do
csp_log_path = Rails.root.join("log", "csp_violations.log")
logger = Logger.new(
csp_log_path,
'daily', # Rotate daily
"daily", # Rotate daily
30 # Keep 30 old log files
)
logger.level = Logger::INFO
logger.formatter = proc do |severity, datetime, progname, msg|
"[#{datetime.strftime('%Y-%m-%d %H:%M:%S')}] #{severity} #{msg}\n"
"[#{datetime.strftime("%Y-%m-%d %H:%M:%S")}] #{severity} #{msg}\n"
end
logger
end
@@ -120,7 +119,6 @@ Rails.application.config.after_initialize do
# Test write to ensure permissions are correct
csp_logger.info "CSP Logger initialized at #{Time.current}"
rescue => e
Rails.logger.error "Failed to initialize CSP local logger: #{e.message}"
Rails.logger.error "CSP violations will only be sent to Sentry (if configured)"

View File

@@ -4,5 +4,5 @@
# Use this to limit dissemination of sensitive information.
# See the ActiveSupport::ParameterFilter documentation for supported notations and behaviors.
Rails.application.config.filter_parameters += [
:passw, :email, :secret, :token, :_key, :crypt, :salt, :certificate, :otp, :ssn, :cvv, :cvc
:passw, :email, :secret, :token, :_key, :crypt, :salt, :certificate, :otp, :ssn, :cvv, :cvc, :backup
]

View File

@@ -0,0 +1,19 @@
# Configure the Permissions-Policy header
# See https://api.rubyonrails.org/classes/ActionDispatch/PermissionsPolicy.html
Rails.application.config.permissions_policy do |f|
# Disable sensitive browser features for security
f.camera :none
f.gyroscope :none
f.microphone :none
f.payment :none
f.usb :none
f.magnetometer :none
# You can enable specific features as needed:
# f.fullscreen :self
# f.geolocation :self
# You can also allow specific origins:
# f.payment :self, "https://secure.example.com"
end

View File

@@ -47,7 +47,7 @@ Rails.application.config.after_initialize do
timestamp: csp_data[:timestamp]
}
},
user: csp_data[:current_user_id] ? { id: csp_data[:current_user_id] } : nil
user: csp_data[:current_user_id] ? {id: csp_data[:current_user_id]} : nil
)
# Log to Rails logger for redundancy
@@ -69,10 +69,10 @@ Rails.application.config.after_initialize do
parsed.host
rescue URI::InvalidURIError
# Handle cases where URI might be malformed or just a path
if uri.start_with?('/')
if uri.start_with?("/")
nil # It's a relative path, no domain
else
uri.split('/').first # Best effort extraction
uri.split("/").first # Best effort extraction
end
end
end

View File

@@ -0,0 +1,7 @@
# Token HMAC key derivation
# This key is used to compute HMAC-based token prefixes for fast lookup
# Derived from SECRET_KEY_BASE - no storage needed, deterministic output
# Optional: Set OIDC_TOKEN_PREFIX_HMAC env var to override with explicit key
module TokenHmac
KEY = ENV["OIDC_TOKEN_PREFIX_HMAC"] || Rails.application.key_generator.generate_key("oidc_token_prefix", 32)
end

View File

@@ -0,0 +1,5 @@
# frozen_string_literal: true
module Clinch
VERSION = "0.8.4"
end

View File

@@ -31,7 +31,6 @@ threads threads_count, threads_count
# Specifies the `port` that Puma will listen on to receive requests; default is 3000.
port ENV.fetch("PORT", 3000)
# Allow puma to be restarted by `bin/rails restart` command.
plugin :tmp_restart

View File

@@ -8,7 +8,7 @@ Rails.application.routes.draw do
# Reveal health status on /up that returns 200 if the app boots with no exceptions, otherwise 500.
# Can be used by load balancers and uptime monitors to verify that the app is live.
get "up" => "rails/health#show", as: :rails_health_check
get "up" => "rails/health#show", :as => :rails_health_check
# Authentication routes
get "/signup", to: "users#new", as: :signup
@@ -49,6 +49,7 @@ Rails.application.routes.draw do
end
resource :active_sessions, only: [:show] do
member do
delete :logout_from_app
delete :revoke_consent
delete :revoke_all_consents
end
@@ -60,21 +61,21 @@ Rails.application.routes.draw do
end
# TOTP (2FA) routes
get '/totp/new', to: 'totp#new', as: :new_totp
post '/totp', to: 'totp#create', as: :totp
delete '/totp', to: 'totp#destroy'
get '/totp/backup_codes', to: 'totp#backup_codes', as: :backup_codes_totp
post '/totp/verify_password', to: 'totp#verify_password', as: :verify_password_totp
get '/totp/regenerate_backup_codes', to: 'totp#regenerate_backup_codes', as: :regenerate_backup_codes_totp
post '/totp/regenerate_backup_codes', to: 'totp#create_new_backup_codes', as: :create_new_backup_codes_totp
post '/totp/complete_setup', to: 'totp#complete_setup', as: :complete_totp_setup
get "/totp/new", to: "totp#new", as: :new_totp
post "/totp", to: "totp#create", as: :totp
delete "/totp", to: "totp#destroy"
get "/totp/backup_codes", to: "totp#backup_codes", as: :backup_codes_totp
post "/totp/verify_password", to: "totp#verify_password", as: :verify_password_totp
get "/totp/regenerate_backup_codes", to: "totp#regenerate_backup_codes", as: :regenerate_backup_codes_totp
post "/totp/regenerate_backup_codes", to: "totp#create_new_backup_codes", as: :create_new_backup_codes_totp
post "/totp/complete_setup", to: "totp#complete_setup", as: :complete_totp_setup
# WebAuthn (Passkeys) routes
get '/webauthn/new', to: 'webauthn#new', as: :new_webauthn
post '/webauthn/challenge', to: 'webauthn#challenge'
post '/webauthn/create', to: 'webauthn#create'
delete '/webauthn/:id', to: 'webauthn#destroy', as: :webauthn_credential
get '/webauthn/check', to: 'webauthn#check'
get "/webauthn/new", to: "webauthn#new", as: :new_webauthn
post "/webauthn/challenge", to: "webauthn#challenge"
post "/webauthn/create", to: "webauthn#create"
delete "/webauthn/:id", to: "webauthn#destroy", as: :webauthn_credential
get "/webauthn/check", to: "webauthn#check"
# Admin routes
namespace :admin do

View File

@@ -4,7 +4,7 @@ test:
local:
service: Disk
root: <%= Rails.root.join("storage") %>
root: <%= Rails.root.join("storage/uploads") %>
# Use bin/rails credentials:edit to set the AWS secrets (as aws:access_key_id|secret_access_key)
# amazon:

View File

@@ -7,6 +7,6 @@ class CreateUserGroups < ActiveRecord::Migration[8.1]
t.timestamps
end
add_index :user_groups, [ :user_id, :group_id ], unique: true
add_index :user_groups, [:user_id, :group_id], unique: true
end
end

View File

@@ -7,6 +7,6 @@ class CreateApplicationGroups < ActiveRecord::Migration[8.1]
t.timestamps
end
add_index :application_groups, [ :application_id, :group_id ], unique: true
add_index :application_groups, [:application_id, :group_id], unique: true
end
end

View File

@@ -13,6 +13,6 @@ class CreateOidcAuthorizationCodes < ActiveRecord::Migration[8.1]
end
add_index :oidc_authorization_codes, :code, unique: true
add_index :oidc_authorization_codes, :expires_at
add_index :oidc_authorization_codes, [ :application_id, :user_id ]
add_index :oidc_authorization_codes, [:application_id, :user_id]
end
end

View File

@@ -11,6 +11,6 @@ class CreateOidcAccessTokens < ActiveRecord::Migration[8.1]
end
add_index :oidc_access_tokens, :token, unique: true
add_index :oidc_access_tokens, :expires_at
add_index :oidc_access_tokens, [ :application_id, :user_id ]
add_index :oidc_access_tokens, [:application_id, :user_id]
end
end

View File

@@ -1,9 +1,9 @@
class AddRoleMappingToApplications < ActiveRecord::Migration[8.1]
def change
add_column :applications, :role_mapping_mode, :string, default: 'disabled', null: false
add_column :applications, :role_mapping_mode, :string, default: "disabled", null: false
add_column :applications, :role_prefix, :string
add_column :applications, :managed_permissions, :json, default: {}
add_column :applications, :role_claim_name, :string, default: 'roles'
add_column :applications, :role_claim_name, :string, default: "roles"
create_table :application_roles do |t|
t.references :application, null: false, foreign_key: true
@@ -21,7 +21,7 @@ class AddRoleMappingToApplications < ActiveRecord::Migration[8.1]
create_table :user_role_assignments do |t|
t.references :user, null: false, foreign_key: true
t.references :application_role, null: false, foreign_key: true
t.string :source, default: 'oidc' # 'oidc', 'manual', 'group_sync'
t.string :source, default: "oidc" # 'oidc', 'manual', 'group_sync'
t.json :metadata, default: {}
t.timestamps

View File

@@ -41,7 +41,7 @@ class MigrateForwardAuthRulesToApplications < ActiveRecord::Migration[8.1]
app = application_class.create!(
name: rule.domain_pattern.titleize,
slug: rule.domain_pattern.parameterize.presence || "forward-auth-#{rule.id}",
app_type: 'forward_auth',
app_type: "forward_auth",
domain_pattern: rule.domain_pattern,
headers_config: rule.headers_config || {},
active: rule.active
@@ -59,7 +59,7 @@ class MigrateForwardAuthRulesToApplications < ActiveRecord::Migration[8.1]
def down
# Remove all forward_auth applications created by this migration
Application.where(app_type: 'forward_auth').destroy_all
Application.where(app_type: "forward_auth").destroy_all
end
private

View File

@@ -5,7 +5,7 @@ class CreateWebauthnCredentials < ActiveRecord::Migration[8.1]
t.references :user, null: false, foreign_key: true, index: true
# WebAuthn specification fields
t.string :external_id, null: false, index: { unique: true } # credential ID (base64)
t.string :external_id, null: false, index: {unique: true} # credential ID (base64)
t.string :public_key, null: false # public key (base64)
t.integer :sign_count, null: false, default: 0 # signature counter (clone detection)

View File

@@ -17,6 +17,6 @@ class CreateOidcRefreshTokens < ActiveRecord::Migration[8.1]
add_index :oidc_refresh_tokens, :expires_at
add_index :oidc_refresh_tokens, :revoked_at
add_index :oidc_refresh_tokens, :token_family_id
add_index :oidc_refresh_tokens, [ :application_id, :user_id ]
add_index :oidc_refresh_tokens, [:application_id, :user_id]
end
end

View File

@@ -1,13 +1,13 @@
class CreateApplicationUserClaims < ActiveRecord::Migration[8.1]
def change
create_table :application_user_claims do |t|
t.references :application, null: false, foreign_key: { on_delete: :cascade }
t.references :user, null: false, foreign_key: { on_delete: :cascade }
t.references :application, null: false, foreign_key: {on_delete: :cascade}
t.references :user, null: false, foreign_key: {on_delete: :cascade}
t.json :custom_claims, default: {}, null: false
t.timestamps
end
add_index :application_user_claims, [:application_id, :user_id], unique: true, name: 'index_app_user_claims_unique'
add_index :application_user_claims, [:application_id, :user_id], unique: true, name: "index_app_user_claims_unique"
end
end

View File

@@ -19,7 +19,7 @@ class CreateActiveStorageTables < ActiveRecord::Migration[7.0]
t.datetime :created_at, null: false
end
t.index [ :key ], unique: true
t.index [:key], unique: true
end
create_table :active_storage_attachments, id: primary_key_type do |t|
@@ -33,7 +33,7 @@ class CreateActiveStorageTables < ActiveRecord::Migration[7.0]
t.datetime :created_at, null: false
end
t.index [ :record_type, :record_id, :name, :blob_id ], name: :index_active_storage_attachments_uniqueness, unique: true
t.index [:record_type, :record_id, :name, :blob_id], name: :index_active_storage_attachments_uniqueness, unique: true
t.foreign_key :active_storage_blobs, column: :blob_id
end
@@ -41,17 +41,18 @@ class CreateActiveStorageTables < ActiveRecord::Migration[7.0]
t.belongs_to :blob, null: false, index: false, type: foreign_key_type
t.string :variation_digest, null: false
t.index [ :blob_id, :variation_digest ], name: :index_active_storage_variant_records_uniqueness, unique: true
t.index [:blob_id, :variation_digest], name: :index_active_storage_variant_records_uniqueness, unique: true
t.foreign_key :active_storage_blobs, column: :blob_id
end
end
private
def primary_and_foreign_key_types
config = Rails.configuration.generators
setting = config.options[config.orm][:primary_key_type]
primary_key_type = setting || :primary_key
foreign_key_type = setting || :bigint
[ primary_key_type, foreign_key_type ]
[primary_key_type, foreign_key_type]
end
end

View File

@@ -0,0 +1,5 @@
class AddBackchannelLogoutUriToApplications < ActiveRecord::Migration[8.1]
def change
add_column :applications, :backchannel_logout_uri, :string
end
end

View File

@@ -0,0 +1,42 @@
class AddTokenPrefixToTokens < ActiveRecord::Migration[8.1]
def up
add_column :oidc_access_tokens, :token_prefix, :string, limit: 8
add_column :oidc_refresh_tokens, :token_prefix, :string, limit: 8
# Backfill existing tokens with prefix and digest
say_with_time "Backfilling token prefixes and digests..." do
[OidcAccessToken, OidcRefreshToken].each do |klass|
klass.reset_column_information # Ensure Rails knows about new column
klass.where(token_prefix: nil).find_each do |token|
next unless token.token.present?
updates = {}
# Compute HMAC prefix
prefix = klass.compute_token_prefix(token.token)
updates[:token_prefix] = prefix if prefix.present?
# Backfill digest if missing
if token.token_digest.nil?
updates[:token_digest] = BCrypt::Password.create(token.token)
end
token.update_columns(updates) if updates.any?
end
say " #{klass.name}: #{klass.where.not(token_prefix: nil).count} tokens backfilled"
end
end
add_index :oidc_access_tokens, :token_prefix
add_index :oidc_refresh_tokens, :token_prefix
end
def down
remove_index :oidc_access_tokens, :token_prefix
remove_index :oidc_refresh_tokens, :token_prefix
remove_column :oidc_access_tokens, :token_prefix
remove_column :oidc_refresh_tokens, :token_prefix
end
end

View File

@@ -0,0 +1,10 @@
class RemovePlaintextTokenFromOidcAccessTokens < ActiveRecord::Migration[8.1]
def change
# Remove the unique index first
remove_index :oidc_access_tokens, :token, if_exists: true
# Remove the plaintext token column - no longer needed
# Tokens are now stored as BCrypt-hashed token_digest with HMAC token_prefix
remove_column :oidc_access_tokens, :token, :string
end
end

View File

@@ -0,0 +1,14 @@
class AddPkceOptionsToApplications < ActiveRecord::Migration[8.1]
def change
# Add require_pkce column for confidential clients
# Default true for new apps (secure by default), existing apps will be false
add_column :applications, :require_pkce, :boolean, default: true, null: false
# Set existing applications to not require PKCE (backwards compatibility)
reversible do |dir|
dir.up do
execute "UPDATE applications SET require_pkce = false WHERE id > 0"
end
end
end
end

View File

@@ -0,0 +1,20 @@
class RenameCodeToCodeHmacAndAddTokenHmac < ActiveRecord::Migration[8.1]
def change
# Authorization codes: rename code to code_hmac
rename_column :oidc_authorization_codes, :code, :code_hmac
# Access tokens: add token_hmac, remove old columns
add_column :oidc_access_tokens, :token_hmac, :string
add_index :oidc_access_tokens, :token_hmac, unique: true
remove_column :oidc_access_tokens, :token_prefix
remove_column :oidc_access_tokens, :token_digest
# Refresh tokens: add token_hmac, remove old columns
add_column :oidc_refresh_tokens, :token_hmac, :string
add_index :oidc_refresh_tokens, :token_hmac, unique: true
remove_column :oidc_refresh_tokens, :token_prefix
remove_column :oidc_refresh_tokens, :token_digest
end
end

View File

@@ -0,0 +1,6 @@
class AddAuthTimeToOidcTokens < ActiveRecord::Migration[8.1]
def change
add_column :oidc_authorization_codes, :auth_time, :integer
add_column :oidc_refresh_tokens, :auth_time, :integer
end
end

View File

@@ -0,0 +1,7 @@
class AddAcrToOidcTokensAndSessions < ActiveRecord::Migration[8.1]
def change
add_column :sessions, :acr, :string
add_column :oidc_authorization_codes, :acr, :string
add_column :oidc_refresh_tokens, :acr, :string
end
end

53
db/schema.rb generated
View File

@@ -10,7 +10,35 @@
#
# It's strongly recommended that you check this file into your version control system.
ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
ActiveRecord::Schema[8.1].define(version: 2025_12_31_060112) do
create_table "active_storage_attachments", force: :cascade do |t|
t.bigint "blob_id", null: false
t.datetime "created_at", null: false
t.string "name", null: false
t.bigint "record_id", null: false
t.string "record_type", null: false
t.index ["blob_id"], name: "index_active_storage_attachments_on_blob_id"
t.index ["record_type", "record_id", "name", "blob_id"], name: "index_active_storage_attachments_uniqueness", unique: true
end
create_table "active_storage_blobs", force: :cascade do |t|
t.bigint "byte_size", null: false
t.string "checksum"
t.string "content_type"
t.datetime "created_at", null: false
t.string "filename", null: false
t.string "key", null: false
t.text "metadata"
t.string "service_name", null: false
t.index ["key"], name: "index_active_storage_blobs_on_key", unique: true
end
create_table "active_storage_variant_records", force: :cascade do |t|
t.bigint "blob_id", null: false
t.string "variation_digest", null: false
t.index ["blob_id", "variation_digest"], name: "index_active_storage_variant_records_uniqueness", unique: true
end
create_table "application_groups", force: :cascade do |t|
t.integer "application_id", null: false
t.datetime "created_at", null: false
@@ -36,6 +64,7 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
t.integer "access_token_ttl", default: 3600
t.boolean "active", default: true, null: false
t.string "app_type", null: false
t.string "backchannel_logout_uri"
t.string "client_id"
t.string "client_secret_digest"
t.datetime "created_at", null: false
@@ -48,6 +77,7 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
t.string "name", null: false
t.text "redirect_uris"
t.integer "refresh_token_ttl", default: 2592000
t.boolean "require_pkce", default: true, null: false
t.string "slug", null: false
t.datetime "updated_at", null: false
t.index ["active"], name: "index_applications_on_active"
@@ -71,24 +101,24 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
t.datetime "expires_at", null: false
t.datetime "revoked_at"
t.string "scope"
t.string "token"
t.string "token_digest"
t.string "token_hmac"
t.datetime "updated_at", null: false
t.integer "user_id", null: false
t.index ["application_id", "user_id"], name: "index_oidc_access_tokens_on_application_id_and_user_id"
t.index ["application_id"], name: "index_oidc_access_tokens_on_application_id"
t.index ["expires_at"], name: "index_oidc_access_tokens_on_expires_at"
t.index ["revoked_at"], name: "index_oidc_access_tokens_on_revoked_at"
t.index ["token"], name: "index_oidc_access_tokens_on_token", unique: true
t.index ["token_digest"], name: "index_oidc_access_tokens_on_token_digest", unique: true
t.index ["token_hmac"], name: "index_oidc_access_tokens_on_token_hmac", unique: true
t.index ["user_id"], name: "index_oidc_access_tokens_on_user_id"
end
create_table "oidc_authorization_codes", force: :cascade do |t|
t.string "acr"
t.integer "application_id", null: false
t.string "code", null: false
t.integer "auth_time"
t.string "code_challenge"
t.string "code_challenge_method"
t.string "code_hmac", null: false
t.datetime "created_at", null: false
t.datetime "expires_at", null: false
t.string "nonce"
@@ -99,21 +129,23 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
t.integer "user_id", null: false
t.index ["application_id", "user_id"], name: "index_oidc_authorization_codes_on_application_id_and_user_id"
t.index ["application_id"], name: "index_oidc_authorization_codes_on_application_id"
t.index ["code"], name: "index_oidc_authorization_codes_on_code", unique: true
t.index ["code_challenge"], name: "index_oidc_authorization_codes_on_code_challenge"
t.index ["code_hmac"], name: "index_oidc_authorization_codes_on_code_hmac", unique: true
t.index ["expires_at"], name: "index_oidc_authorization_codes_on_expires_at"
t.index ["user_id"], name: "index_oidc_authorization_codes_on_user_id"
end
create_table "oidc_refresh_tokens", force: :cascade do |t|
t.string "acr"
t.integer "application_id", null: false
t.integer "auth_time"
t.datetime "created_at", null: false
t.datetime "expires_at", null: false
t.integer "oidc_access_token_id", null: false
t.datetime "revoked_at"
t.string "scope"
t.string "token_digest", null: false
t.integer "token_family_id"
t.string "token_hmac"
t.datetime "updated_at", null: false
t.integer "user_id", null: false
t.index ["application_id", "user_id"], name: "index_oidc_refresh_tokens_on_application_id_and_user_id"
@@ -121,8 +153,8 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
t.index ["expires_at"], name: "index_oidc_refresh_tokens_on_expires_at"
t.index ["oidc_access_token_id"], name: "index_oidc_refresh_tokens_on_oidc_access_token_id"
t.index ["revoked_at"], name: "index_oidc_refresh_tokens_on_revoked_at"
t.index ["token_digest"], name: "index_oidc_refresh_tokens_on_token_digest", unique: true
t.index ["token_family_id"], name: "index_oidc_refresh_tokens_on_token_family_id"
t.index ["token_hmac"], name: "index_oidc_refresh_tokens_on_token_hmac", unique: true
t.index ["user_id"], name: "index_oidc_refresh_tokens_on_user_id"
end
@@ -142,6 +174,7 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
end
create_table "sessions", force: :cascade do |t|
t.string "acr"
t.datetime "created_at", null: false
t.string "device_name"
t.datetime "expires_at"
@@ -211,6 +244,8 @@ ActiveRecord::Schema[8.1].define(version: 2025_11_25_012446) do
t.index ["user_id"], name: "index_webauthn_credentials_on_user_id"
end
add_foreign_key "active_storage_attachments", "active_storage_blobs", column: "blob_id"
add_foreign_key "active_storage_variant_records", "active_storage_blobs", column: "blob_id"
add_foreign_key "application_groups", "applications"
add_foreign_key "application_groups", "groups"
add_foreign_key "application_user_claims", "applications", on_delete: :cascade

316
docs/backchannel-logout.md Normal file
View File

@@ -0,0 +1,316 @@
# OpenID Connect Backchannel Logout
## Overview
Backchannel logout is an OpenID Connect feature that enables Clinch to notify applications when a user logs out, ensuring sessions are terminated across all connected applications immediately.
## How It Works
When a user logs out from Clinch (or any connected application), Clinch sends server-to-server HTTP POST requests to all applications that have configured a backchannel logout endpoint. This happens automatically in the background.
### Logout Triggers
Backchannel logout notifications are sent when:
1. **User clicks "Sign Out" in Clinch** - All connected OIDC applications are notified, then the Clinch session is terminated
2. **User logs out via OIDC `/logout` endpoint** (RP-Initiated Logout) - All connected applications are notified, then the Clinch session is terminated
3. **User clicks "Logout" on an app (Dashboard)** - Backchannel logout is sent to that app, all access/refresh tokens are revoked, but OAuth consent is preserved (user can sign back in without re-authorizing)
4. **User clicks "Revoke Access" for a specific app (Active Sessions page)** - Backchannel logout is sent to that app to terminate its session, all access/refresh tokens are revoked, then the OAuth consent is permanently destroyed (user must re-authorize the app to use it again)
5. **User clicks "Revoke All App Access"** - All connected applications receive backchannel logout notifications, all tokens are revoked, then all OAuth consents are permanently destroyed
### The Logout Flow
```
User logs out → Clinch finds all connected apps
For each app with backchannel_logout_uri:
Generate signed JWT logout token
HTTP POST to app's logout endpoint
App validates JWT and terminates session
Clinch revokes access and refresh tokens
```
### Logout vs Revoke Access
Clinch provides two distinct actions for managing application access:
| Action | Location | What Happens | When to Use |
|--------|----------|--------------|-------------|
| **Logout** | Dashboard | • Sends backchannel logout to app<br>• Revokes all access tokens<br>• Revokes all refresh tokens<br>• **Keeps OAuth consent intact** | You want to end your session with an app but still trust it. Next login will skip the authorization screen. |
| **Revoke Access** | Active Sessions page | • Sends backchannel logout to app<br>• Revokes all access tokens<br>• Revokes all refresh tokens<br>• **Destroys OAuth consent** | You want to completely de-authorize an app. Next login will require you to re-authorize the app. |
**Key Difference**: "Logout" preserves the authorization relationship while terminating the active session. "Revoke Access" completely removes the app's authorization to access your account.
**Example Use Cases**:
- **Logout**: "I left my Jellyfin session open at a friend's house. I want to kill that session but I still use Jellyfin."
- **Revoke Access**: "I no longer trust this app and want to remove its authorization completely."
**Technical Details**:
- Both actions revoke access tokens (opaque, database-backed, validated on each use)
- Both actions revoke refresh tokens (prevents obtaining new access tokens)
- ID tokens remain valid until expiry (stateless JWTs), but apps should honor backchannel logout
- Backchannel logout ensures the app clears its local session immediately
## Configuring Applications
### In Clinch Admin UI
1. Navigate to **Admin → Applications**
2. Edit or create an OIDC application
3. In the "Backchannel Logout URI" field, enter the application's logout endpoint
- Example: `https://kavita.local/oidc/backchannel-logout`
- Must be HTTPS in production
- Leave blank if the application doesn't support backchannel logout
### Checking Support
The OIDC discovery endpoint advertises backchannel logout support:
```bash
curl https://clinch.local/.well-known/openid-configuration | jq
```
Look for:
```json
{
"backchannel_logout_supported": true,
"backchannel_logout_session_supported": true
}
```
## Implementing a Backchannel Logout Endpoint (for RPs)
If you're developing an application that integrates with Clinch, here's how to implement backchannel logout support:
### 1. Create the Endpoint
The endpoint must:
- Accept HTTP POST requests
- Parse the `logout_token` parameter from the form body
- Validate the JWT signature
- Terminate the user's session
- Return 200 OK quickly (within 5 seconds)
### 2. Example Implementation (Ruby/Rails)
```ruby
# config/routes.rb
post '/oidc/backchannel-logout', to: 'oidc_backchannel_logout#logout'
# app/controllers/oidc_backchannel_logout_controller.rb
class OidcBackchannelLogoutController < ApplicationController
skip_before_action :verify_authenticity_token # Server-to-server call
skip_before_action :authenticate_user! # No user session yet
def logout
logout_token = params[:logout_token]
unless logout_token.present?
head :bad_request
return
end
begin
# Decode and verify the JWT
# Get Clinch's public key from JWKS endpoint
jwks = fetch_clinch_jwks
decoded = JWT.decode(
logout_token,
nil, # Will be verified using JWKS
true,
{
algorithms: ['RS256'],
jwks: jwks,
verify_aud: true,
aud: YOUR_CLIENT_ID,
verify_iss: true,
iss: 'https://clinch.local' # Your Clinch URL
}
)
claims = decoded.first
# Validate required claims
unless claims['events']&.key?('http://schemas.openid.net/event/backchannel-logout')
head :bad_request
return
end
# Get session ID from the token
sid = claims['sid']
sub = claims['sub']
# Terminate sessions
if sid.present?
# Terminate specific session by SID (recommended)
Session.where(oidc_sid: sid).destroy_all
elsif sub.present?
# Terminate all sessions for this user
user = User.find_by(oidc_sub: sub)
user&.sessions&.destroy_all
end
Rails.logger.info "Backchannel logout: Terminated session for sid=#{sid}, sub=#{sub}"
head :ok
rescue JWT::DecodeError => e
Rails.logger.error "Backchannel logout: Invalid JWT - #{e.message}"
head :bad_request
rescue => e
Rails.logger.error "Backchannel logout: Error - #{e.class}: #{e.message}"
head :internal_server_error
end
end
private
def fetch_clinch_jwks
# Cache this in production!
response = HTTParty.get('https://clinch.local/.well-known/jwks.json')
JSON.parse(response.body, symbolize_names: true)
end
end
```
### 3. Required JWT Claims Validation
The logout token will contain:
| Claim | Description | Required |
|-------|-------------|----------|
| `iss` | Issuer (Clinch URL) | Yes |
| `aud` | Your application's client_id | Yes |
| `iat` | Issued at timestamp | Yes |
| `jti` | Unique token ID | Yes |
| `sub` | Pairwise subject identifier (user's SID) | Yes |
| `sid` | Session ID (same as sub) | Yes |
| `events` | Must contain `http://schemas.openid.net/event/backchannel-logout` | Yes |
| `nonce` | Must NOT be present (spec requirement) | No |
### 4. Session Tracking Requirements
To support backchannel logout, your application must:
1. **Store the `sid` claim from ID tokens**:
```ruby
# When user logs in via OIDC
id_token = decode_id_token(params[:id_token])
session[:oidc_sid] = id_token['sid'] # Store this!
```
2. **Associate sessions with SID**:
```ruby
# Create session with SID tracking
Session.create!(
user: current_user,
oidc_sid: id_token['sid'],
...
)
```
3. **Terminate sessions by SID**:
```ruby
# When backchannel logout is received
Session.where(oidc_sid: sid).destroy_all
```
### 5. Testing Your Endpoint
Test with curl:
```bash
# Get a valid logout token (you'll need to capture this from Clinch logs)
LOGOUT_TOKEN="eyJhbGc..."
curl -X POST https://your-app.local/oidc/backchannel-logout \
-H "Content-Type: application/x-www-form-urlencoded" \
-d "logout_token=$LOGOUT_TOKEN"
```
Expected response: `200 OK` (empty body)
## Monitoring and Troubleshooting
### Checking Logs
Clinch logs all backchannel logout attempts:
```bash
# In development
tail -f log/development.log | grep BackchannelLogout
# Example log output:
# BackchannelLogout: Successfully sent logout notification to Kavita (https://kavita.local/oidc/backchannel-logout)
# BackchannelLogout: Application Jellyfin doesn't support backchannel logout
# BackchannelLogout: Timeout sending logout to HomeAssistant (https://ha.local/logout): Connection timeout
```
### Common Issues
**1. HTTP Timeout**
- Symptom: `Timeout sending logout to...` in logs
- Solution: Ensure the RP's backchannel logout endpoint responds within 5 seconds
- Note: Clinch will retry 3 times with exponential backoff
**2. HTTP Errors (Non-200 Status)**
- Symptom: `Application X returned HTTP 400/500...` in logs
- Solution: Check the RP's logs for JWT validation errors
- Common causes:
- Wrong JWKS (public key mismatch)
- Incorrect `aud` (client_id) validation
- Missing required claims validation
**3. Network Unreachable**
- Symptom: `Failed to send logout to...` with connection errors
- Solution: Ensure the RP's logout endpoint is accessible from Clinch server
- Check: Firewalls, DNS, SSL certificates
**4. Sessions Not Terminating**
- Symptom: User still logged into RP after logging out of Clinch
- Solution: Verify the RP is storing and checking `sid` correctly
- Debug: Add logging to the RP's backchannel logout handler
### Verification Checklist
For RPs (Application Developers):
- [ ] Endpoint accepts POST requests
- [ ] Endpoint validates JWT signature using Clinch's JWKS
- [ ] Endpoint validates all required claims
- [ ] Endpoint terminates sessions by SID
- [ ] Endpoint returns 200 OK quickly (< 5 seconds)
- [ ] Sessions store the `sid` claim from ID tokens
- [ ] Backchannel logout URI is configured in Clinch admin
For Administrators:
- [ ] Application has `backchannel_logout_uri` configured
- [ ] URI uses HTTPS (in production)
- [ ] URI is reachable from Clinch server
- [ ] Check logs for successful logout notifications
## Security Considerations
1. **JWT Signature Verification**: Always verify the logout token signature using Clinch's public key
2. **Audience Validation**: Ensure the `aud` claim matches your client_id
3. **Issuer Validation**: Ensure the `iss` claim matches your Clinch URL
4. **No Authentication Required**: The endpoint should not require user authentication (it's server-to-server)
5. **HTTPS Only**: Always use HTTPS in production (Clinch enforces this)
6. **Fire-and-Forget**: RPs should log failures but not block on errors
## Comparison with Other Logout Methods
| Method | Communication | When Sessions Terminate | Reliability |
|--------|--------------|------------------------|-------------|
| **Backchannel Logout** | Server-to-server POST | Immediately | High (retries on failure) |
| **Front-Channel Logout** | Browser iframes | When browser loads iframes | Low (blocked by privacy settings) |
| **RP-Initiated Logout** | User redirects to Clinch | Only affects Clinch session | N/A (just triggers other methods) |
| **Token Expiry** | None | When access token expires | Guaranteed but delayed |
## References
- [OpenID Connect Back-Channel Logout 1.0](https://openid.net/specs/openid-connect-backchannel-1_0.html)
- [RFC 7009: OAuth 2.0 Token Revocation](https://tools.ietf.org/html/rfc7009)
- [Clinch OIDC Discovery](/.well-known/openid-configuration)

280
docs/beta-checklist.md Normal file
View File

@@ -0,0 +1,280 @@
# Beta Release Readiness Checklist
This checklist ensures Clinch meets security, quality, and documentation standards before moving from "experimental" to "Beta" status.
> **Security Implementation Status:** See [security-todo.md](security-todo.md) for detailed vulnerability tracking and fixes.
> **Outstanding Security Issues:** 3 (all MEDIUM/LOW priority) - Phases 1-4 complete ✅
---
## Security Scanning
### Automated Security Tools
- [x] **Brakeman** - Static security analysis for Rails
- Status: ✅ Passing (2 weak warnings documented and accepted)
- Command: `bin/brakeman --no-pager`
- CI: Runs on every PR and push to main
- Warnings documented in `config/brakeman.ignore`
- [x] **bundler-audit** - Dependency vulnerability scanning
- Status: ✅ No vulnerabilities found
- Command: `bin/bundler-audit check --update`
- CI: Runs on every PR and push to main
- [x] **importmap audit** - JavaScript dependency scanning
- CI: Runs on every PR and push to main
- [x] **Trivy** - Container image vulnerability scanning
- Scans Docker images for OS and system package vulnerabilities
- CI: Builds and scans image on every PR and push to main
- Results uploaded to GitHub Security tab
- [x] **Dependabot** - Automated dependency updates
- Creates PRs for outdated dependencies
- Enabled for Ruby gems and GitHub Actions
- [x] **GitHub Secret Scanning** - Detects leaked credentials
- Push protection enabled to block commits with secrets
- [x] **Test Coverage** - SimpleCov integration
- Command: `COVERAGE=1 bin/rails test`
- Coverage report: `coverage/index.html`
### Security Features Implemented
#### Authentication
- [x] Secure password storage (bcrypt with Rails defaults)
- [x] TOTP 2FA with backup codes
- [x] WebAuthn/Passkey support (FIDO2)
- [x] Session management with device tracking
- [x] Session revocation (individual and bulk)
- [x] Remember me with configurable expiry
- [x] Account invitation flow with expiring tokens
- [x] Password reset with expiring tokens
#### OIDC Security
- [x] Authorization code flow with PKCE support
- [x] Refresh token rotation
- [x] Token family tracking (detects replay attacks)
- [x] All tokens HMAC-SHA256 hashed in database
- [x] Configurable token expiry (access, refresh, ID)
- [x] One-time use authorization codes
- [x] Pairwise subject identifiers (privacy)
- [x] ID tokens signed with RS256
- [x] Token revocation endpoint (RFC 7009)
- [x] Proper `at_hash` validation
- [x] OIDC standard claims (auth_time, acr, azp)
- [x] Automatic cleanup of expired tokens
#### Access Control
- [x] Group-based authorization
- [x] Application-level access control
- [x] Admin vs. regular user roles
- [x] User status management (active, disabled, pending)
- [x] TOTP enforcement per-user
- [x] ForwardAuth policy enforcement
#### Input Validation
- [x] Strong parameter filtering
- [x] URL validation for redirect URIs and landing URLs
- [x] Email validation and normalization
- [x] Slug validation (alphanumeric + hyphens)
- [x] Domain pattern validation for ForwardAuth
- [x] JSON parsing with error handling
- [x] File upload validation (type, size for app icons)
#### Output Encoding
- [x] HTML escaping by default (Rails 8)
- [x] JSON encoding for API responses
- [x] JWT encoding for ID tokens
- [x] Proper content types for responses
#### Session Security
- [x] Secure, httponly cookies
- [x] SameSite cookie attribute
- [x] Session timeout
- [x] IP and User-Agent tracking
- [x] CSRF protection
#### Cryptography
- [x] SecureRandom for tokens
- [x] bcrypt for passwords
- [x] HMAC-SHA256 for token hashing
- [x] RS256 for JWT signing
- [x] Proper secret management (Rails credentials)
## Testing
### Test Coverage
- [x] **341 tests** across integration, model, controller, service, and system tests
- [x] **1349 assertions**
- [x] **0 failures, 0 errors**
### Test Categories
- [x] Integration tests (invitation flow, forward auth, WebAuthn, session security)
- [x] Model tests (OIDC tokens, users, applications, groups, authorization codes)
- [x] Controller tests (TOTP, sessions, passwords, OIDC flows, input validation)
- [x] Service tests (JWT generation and validation)
- [x] System tests (forward auth, WebAuthn security)
### Security-Critical Test Coverage
- [x] OIDC authorization code flow
- [x] PKCE flow
- [x] Refresh token rotation
- [x] Token replay attack detection
- [x] Access control (group-based)
- [x] Input validation
- [x] Session security
- [x] WebAuthn credential handling
- [x] TOTP validation
## Code Quality
- [x] **RuboCop** - Code style and linting
- Configuration: Rails Omakase
- CI: Runs on every PR and push to main
- [x] **Documentation** - Comprehensive README
- Feature documentation
- Setup instructions
- Configuration guide
- Rails console guide
- API/protocol documentation
## Production Readiness
### Configuration
- [ ] Review all environment variables
- [ ] Document required vs. optional configuration
- [ ] Provide sensible defaults
- [ ] Validate production SMTP configuration
- [x] Ensure OIDC private key generation process is documented
### Database
- [x] Migrations are idempotent
- [x] Indexes on foreign keys
- [x] Proper constraints and validations
- [x] SQLite production-ready (Rails 8)
### Performance
- [ ] Review N+1 queries
- [ ] Add database indexes where needed
- [ ] Test with realistic data volumes
- [ ] Review token cleanup job performance
### Deployment
- [x] Docker support
- [x] Docker Compose example
- [x] Production deployment guide (Docker Compose with .env configuration, upgrading, logs)
- [x] Backup and restore documentation
## Security Hardening
### Headers & CSP
- [x] Content Security Policy (comprehensive policy in config/initializers/content_security_policy.rb)
- [x] X-Frame-Options (DENY in production config)
- [x] X-Content-Type-Options (nosniff - Rails default)
- [x] Referrer-Policy (strict-origin-when-cross-origin in production config)
### Rate Limiting
- [x] Login attempt rate limiting (20/3min on sessions#create)
- [x] TOTP verification rate limiting (10/3min on sessions#verify_totp)
- [x] WebAuthn rate limiting (10/1min on webauthn endpoints, 10/3min on session endpoints)
- [x] Password reset rate limiting (10/3min on request, 10/10min on completion)
- [x] Invitation acceptance rate limiting (10/10min)
- [x] OAuth token endpoint rate limiting (60/1min on token, 30/1min on authorize)
- [x] Backup code rate limiting (5 failed attempts per hour, model-level)
### Secrets Management
- [x] No secrets in code
- [x] Rails credentials for sensitive data
- [ ] Document secret rotation process
- [ ] Document OIDC key rotation process
### Logging & Monitoring
- [x] Sentry integration (optional)
- [x] Parameter filtering configured (passwords, tokens, secrets, backup codes, emails filtered from logs)
- [ ] Audit log for admin actions
## Known Limitations & Risks
### Documented Risks
- [x] Document that ForwardAuth requires same-domain setup
- [ ] Document HTTPS requirement for production
- [ ] Document backup code security (single-use, store securely)
- [ ] Document admin password security requirements
### Future Security Enhancements (Post-Beta)
- [x] Rate limiting on authentication endpoints (comprehensive coverage implemented)
- [ ] Account lockout after N failed attempts (rate limiting provides similar protection)
- [ ] Admin audit logging
- [ ] Security event notifications (email/webhook alerts for suspicious activity)
- [ ] Advanced brute force detection (pattern analysis beyond rate limiting)
- [ ] Suspicious login detection (geolocation, device fingerprinting)
- [ ] IP allowlist/blocklist
## External Security Review
- [ ] Consider bug bounty or security audit
- [ ] Penetration testing for OIDC flows
- [ ] WebAuthn implementation review
- [ ] Token security review
## Documentation for Users
- [ ] Security best practices guide
- [ ] Incident response guide
- [x] Backup and disaster recovery guide
- [ ] Upgrade guide
- [ ] Breaking change policy
## Beta Release Criteria
To move from "experimental" to "Beta", the following must be completed:
**Critical (Required for Beta):**
- [x] All automated security scans passing
- [x] All tests passing
- [x] Core features implemented and tested
- [x] Basic documentation complete
- [x] Backup/restore documentation
- [x] Production deployment guide
- [ ] At least one external security review or penetration test
**Important (Should have for Beta):**
- [x] Rate limiting on auth endpoints
- [x] Security headers configuration documented (CSP, X-Frame-Options, X-Content-Type-Options, Referrer-Policy)
- [x] Known limitations documented (ForwardAuth same-domain requirement in README)
- [ ] Admin audit logging
**Nice to have (Can defer to post-Beta):**
- [ ] Bug bounty program
- [ ] Advanced monitoring/alerting
- [x] Automated security testing in CI beyond brakeman/bundler-audit
- [x] Dependabot (automated dependency updates)
- [x] GitHub Secret Scanning (automatic with push protection enabled)
- [x] Container image scanning (Trivy scans Docker images for OS/system vulnerabilities)
- [ ] DAST/Dynamic testing (OWASP ZAP) - optional for post-Beta
## Status Summary
**Current Status:** Pre-Beta / Experimental
**Strengths:**
- ✅ Comprehensive security tooling in place
- ✅ Strong test coverage (341 tests, 1349 assertions)
- ✅ Modern security features (PKCE, token rotation, WebAuthn)
- ✅ Clean security scans (brakeman, bundler-audit)
- ✅ Well-documented codebase
**Before Beta Release:**
- 🔶 External security review recommended
- 🔶 Admin audit logging (optional)
**Recommendation:** Consider Beta status after:
1. External security review or penetration testing
2. Real-world testing period
---
Last updated: 2026-01-02

176
docs/caddy-example.md Normal file
View File

@@ -0,0 +1,176 @@
# Caddy ForwardAuth Configuration Examples
## Basic Configuration (Protecting MEtube)
Assuming Caddy and Clinch are running in a docker compose, and we can use the sevice name `clinch`. Exterally, assume you're connecting to https://clinch.example.com/
```caddyfile
# Clinch SSO (main authentication server)
clinch.yourdomain.com {
reverse_proxy clinch:3000
}
# MEtube (protected by Clinch)
metube.yourdomain.com {
# Forward authentication to Clinch
forward_auth clinch:3000 {
uri /api/verify
# uri /api/verify?rd=https://clinch.yourdomain.com # Shouldn't need this, the rd value should be sent via headers
copy_headers Remote-User Remote-Email Remote-Groups Remote-Admin
}
# If authentication succeeds, proxy to MEtube
handle {
reverse_proxy * {
to http://<ip-address-of-metube>:8081
header_up X-Real-IP {remote_host}
}
}
}
```
## How It Works
1. User visits `https://metube.yourdomain.com`
2. Caddy makes request to `http://clinch:3000/api/verify passing in the url destination for metueb
3. Clinch checks if user is authenticated and authorized:
- If **200**: Caddy forwards request to MEtube with user headers
- If **302**: User is redirected to clinch.yourdomain.com to login
- If **403**: Access denied
4. User signs into Clinch (with TOTP if enabled or Passkey)
5. Clinch redirects back to MEtube
6. User can now access MEtube!
## Protecting Multiple Applications
```caddyfile
# Clinch SSO
clinch.yourdomain.com {
reverse_proxy clinch:3000
}
# MEtube - Anyone can access (no groups required)
metube.yourdomain.com {
forward_auth clinch:3000 {
uri /api/verify
copy_headers Remote-User Remote-Email Remote-Groups Remote-Admin
}
handle {
reverse_proxy * {
to http://metube:8081
header_up X-Real-IP {remote_host}
}
}
}
# Sonarr - Only "media-managers" group
sonarr.yourdomain.com {
forward_auth clinch:3000 {
uri /api/verify
copy_headers Remote-User Remote-Email Remote-Groups Remote-Admin
}
handle {
reverse_proxy * {
to http://sonarr:8989
header_up X-Real-IP {remote_host}
}
}
}
# Grafana - Only "admins" group
grafana.yourdomain.com {
forward_auth clinch:3000 {
uri /api/verify
copy_headers Remote-User Remote-Email Remote-Groups Remote-Admin
}
handle {
reverse_proxy * {
to http://grafana:3001
header_up X-Real-IP {remote_host}
}
}
}
```
## Setup Steps
### 1. Create Applications in Clinch
Create the Application within Clinch, making sure to set Forward Auth application type
### 2. Update Caddyfile
Add the forward_auth directives shown above.
### 3. Reload Caddy
```bash
caddy reload
```
### 4. Test
Visit https://metube.yourdomain.com - you should be redirected to Clinch login!
## Advanced: Passing Headers to Application
Some applications can use the forwarded headers for user identification:
```caddyfile
metube.yourdomain.com {
forward_auth clinch:3000 {
uri /api/verify
copy_headers Remote-User Remote-Email Remote-Groups Remote-Admin
}
# The headers are automatically passed to the backend
handle {
reverse_proxy * {
to http://metube:8081
header_up X-Real-IP {remote_host}
}
}
}
```
Now MEtube receives these headers with every request:
- `Remote-User`: user@example.com
- `Remote-Email`: user@example.com
- `Remote-Groups`: media-managers,users
- `Remote-Admin`: false
## Troubleshooting
### Users not staying logged in
Ensure your Caddy configuration preserves cookies:
```caddyfile
clinch.yourdomain.com {
reverse_proxy localhost:3000 {
header_up X-Forwarded-Host {host}
header_up X-Forwarded-Proto {scheme}
}
}
```
### Authentication loop
Check that the `/api/verify` endpoint is not itself protected:
- `/api/verify` must be accessible without authentication
- It returns 401/403 for unauthenticated users (this is expected)
### Check Clinch logs
```bash
tail -f log/production.log
```
You'll see ForwardAuth log messages like:
```
ForwardAuth: User user@example.com granted access to metube
ForwardAuth: Unauthorized - No session cookie
```

315
docs/claude-review.md Normal file
View File

@@ -0,0 +1,315 @@
# Clinch - Independent Code Review
**Reviewer:** Claude (Anthropic)
**Review Date:** December 2024
**Codebase Version:** Commit 4f31fad
**Review Type:** Security-focused OIDC/OAuth2 correctness review with full application assessment
---
## Executive Summary
Clinch is a self-hosted identity and SSO portal built with Ruby on Rails. This review examined the complete codebase with particular focus on the OIDC/OAuth2 implementation, comparing it against production-grade reference implementations (Rodauth-OAuth, Authelia, Authentik).
**Overall Assessment: Production-Ready**
The implementation demonstrates solid security practices, proper adherence to OAuth 2.0 and OpenID Connect specifications, and comprehensive test coverage. The codebase is well-structured, readable, and maintainable.
---
## Feature Overview
### Authentication Methods
| Feature | Status | Notes |
|---------|--------|-------|
| Password Authentication | Implemented | bcrypt hashing, rate-limited |
| WebAuthn/Passkeys | Implemented | FIDO2 compliant, clone detection |
| TOTP 2FA | Implemented | With backup codes, admin enforcement |
| Session Management | Implemented | Device tracking, revocation |
### SSO Protocols
| Protocol | Status | Notes |
|----------|--------|-------|
| OpenID Connect | Implemented | Full OIDC Core compliance |
| OAuth 2.0 | Implemented | Authorization Code + Refresh Token grants |
| ForwardAuth | Implemented | Traefik/Caddy/nginx compatible |
### User & Access Management
| Feature | Status | Notes |
|---------|--------|-------|
| User CRUD | Implemented | Invitation flow, status management |
| Group Management | Implemented | With custom claims |
| Application Management | Implemented | OIDC + ForwardAuth types |
| Group-based Access Control | Implemented | Per-application restrictions |
---
## OIDC/OAuth2 Implementation Review
### Specification Compliance
| Specification | Status | Evidence |
|---------------|--------|----------|
| RFC 6749 (OAuth 2.0) | Compliant | Proper auth code flow, client authentication |
| RFC 7636 (PKCE) | Compliant | S256 and plain methods, enforced for public clients |
| RFC 7009 (Token Revocation) | Compliant | Always returns 200 OK, prevents scanning |
| OpenID Connect Core 1.0 | Compliant | All required claims, proper JWT structure |
| OIDC Discovery | Compliant | `.well-known/openid-configuration` |
| OIDC Back-Channel Logout | Compliant | Logout tokens per spec |
### ID Token Claims
The implementation includes all required and recommended OIDC claims:
```
Standard: iss, sub, aud, exp, iat, nonce
Profile: email, email_verified, preferred_username, name
Security: at_hash, auth_time, acr, azp
Custom: groups, plus arbitrary claims from groups/users/apps
```
### Token Security
| Aspect | Implementation | Assessment |
|--------|----------------|------------|
| Authorization Codes | HMAC-SHA256 hashed, 10-min expiry, single-use | Secure |
| Access Tokens | HMAC-SHA256 hashed, configurable TTL, indexed lookup | Secure |
| Refresh Tokens | HMAC-SHA256 hashed, rotation with family tracking | Secure |
| ID Tokens | RS256 signed JWTs | Secure |
### Security Features
1. **Authorization Code Reuse Prevention**
- Pessimistic database locking prevents race conditions
- Code reuse triggers revocation of all tokens from that code
- Location: `oidc_controller.rb:342-364`
2. **Refresh Token Rotation**
- Old refresh tokens revoked on use
- Token family tracking detects stolen token reuse
- Revoked token reuse triggers family-wide revocation
- Location: `oidc_controller.rb:504-513`
3. **PKCE Enforcement**
- Required for all public clients
- Configurable for confidential clients
- Proper S256 challenge verification
- Location: `oidc_controller.rb:749-814`
4. **Pairwise Subject Identifiers**
- Each user gets a unique `sub` per application
- Prevents cross-application user tracking
- Location: `oidc_user_consent.rb:59-61`
---
## Security Assessment
### Strengths
1. **Token Storage Architecture**
- All tokens (auth codes, access, refresh) are HMAC-hashed before storage
- Database compromise does not reveal usable tokens
- O(1) indexed lookup via HMAC (not O(n) iteration)
2. **Rate Limiting**
- Sign-in: 20/3min
- TOTP verification: 10/3min
- Token endpoint: 60/min
- Authorization: 30/min
- WebAuthn enumeration check: 10/min
3. **WebAuthn Implementation**
- Sign count validation (clone detection)
- Backup eligibility tracking
- Platform vs roaming authenticator distinction
- Credential enumeration prevention
4. **TOTP Implementation**
- Encrypted secret storage (ActiveRecord Encryption)
- Backup codes are bcrypt-hashed and single-use
- Admin can enforce TOTP requirement per user
5. **Session Security**
- ACR (Authentication Context Class Reference) tracking
- `acr: "1"` for password-only, `acr: "2"` for 2FA/passkey
- Session activity timestamps
- Remote session revocation
### Attack Mitigations
| Attack Vector | Mitigation |
|---------------|------------|
| Credential Stuffing | Rate limiting, account lockout via status |
| Token Theft | HMAC storage, short-lived access tokens, rotation |
| Session Hijacking | Secure cookies, session binding |
| CSRF | Rails CSRF protection, state parameter validation |
| Open Redirect | Strict redirect_uri validation against registered URIs |
| Authorization Code Injection | PKCE enforcement, redirect_uri binding |
| Refresh Token Replay | Token rotation, family-based revocation |
| User Enumeration | Constant-time responses, rate limiting |
### Areas Reviewed (No Issues Found)
- Redirect URI validation (exact match required)
- Client authentication (bcrypt for secrets)
- Error response handling (no sensitive data leakage in production)
- Logout implementation (backchannel notifications, session cleanup)
- Custom claims handling (reserved claim protection)
---
## Code Quality Assessment
### Architecture
| Aspect | Assessment |
|--------|------------|
| Controller Structure | Clean separation, ~900 lines for OIDC (acceptable) |
| Model Design | Well-normalized, proper associations |
| Service Objects | Used appropriately (OidcJwtService, ClaimsMerger) |
| Concerns | TokenPrefixable for shared hashing logic |
### Code Metrics
```
Controllers: ~1,500 lines
Models: ~1,500 lines
Services: ~200 lines
Total App Code: ~3,100 lines
Test Files: 36 files
```
### Readability
- Clear method naming
- Inline documentation for complex logic
- Consistent Ruby style
- No deeply nested conditionals
---
## Test Coverage
### Test Statistics
```
Total Tests: 341
Assertions: 1,349
Failures: 0
Errors: 0
Run Time: 23.5 seconds (parallel)
```
### Test Categories
| Category | Files | Coverage |
|----------|-------|----------|
| OIDC Security | 2 | Auth code reuse, token rotation, PKCE |
| Integration | 4 | WebAuthn, sessions, invitations, forward auth |
| Controllers | 8 | All major endpoints |
| Models | 10 | Validations, associations, business logic |
| Jobs | 4 | Mailers, token cleanup |
### Security-Specific Tests
The test suite includes dedicated security tests:
- `oidc_authorization_code_security_test.rb` - Code reuse, timing attacks, client auth
- `oidc_pkce_controller_test.rb` - PKCE flow validation
- `webauthn_credential_enumeration_test.rb` - Enumeration prevention
- `session_security_test.rb` - Session handling
- `totp_security_test.rb` - 2FA bypass prevention
- `input_validation_test.rb` - Input sanitization
---
## Comparison with Reference Implementations
### vs. Rodauth-OAuth (OpenID Certified)
| Aspect | Rodauth | Clinch |
|--------|---------|--------|
| Modularity | 46 feature modules | Monolithic controller |
| Token Storage | Optional hashing | HMAC-SHA256 (always) |
| PKCE | Dedicated feature | Integrated |
| Certification | OpenID Certified | Not certified |
Clinch has comparable security but less modularity.
### vs. Authelia (Production-Grade Go)
| Aspect | Authelia | Clinch |
|--------|----------|--------|
| PKCE Config | `always/public/never` | Per-app toggle |
| Key Rotation | Supported | Single key |
| PAR Support | Yes | No |
| DPoP Support | Yes | No |
Clinch lacks some advanced features but covers core use cases.
### vs. Authentik (Enterprise Python)
| Aspect | Authentik | Clinch |
|--------|-----------|--------|
| Scale | Enterprise/distributed | Single instance |
| Protocols | OAuth, SAML, LDAP, RADIUS | OAuth/OIDC, ForwardAuth |
| Complexity | High | Low |
Clinch is intentionally simpler for self-hosting.
---
## Recommendations
### Implemented During Review
The following issues were identified and fixed during this review:
1. **Token lookup performance** - Changed from O(n) BCrypt iteration to O(1) HMAC lookup
2. **`at_hash` claim** - Added per OIDC Core spec
3. **`auth_time` claim** - Added for authentication timestamp
4. **`acr` claim** - Added for authentication context class
5. **`azp` claim** - Added for authorized party
6. **Authorization code hashing** - Changed from plaintext to HMAC
7. **Consent SID preservation** - Fixed to preserve pairwise subject ID
8. **Discovery metadata** - Fixed `subject_types_supported` to `["pairwise"]`
### Optional Future Enhancements
| Enhancement | Priority | Effort |
|-------------|----------|--------|
| Key Rotation (multi-key JWKS) | Medium | Medium |
| Token Introspection (RFC 7662) | Low | Low |
| PAR (RFC 9126) | Low | Medium |
| OpenID Certification | Low | High |
---
## Conclusion
Clinch provides a solid, security-conscious OIDC/OAuth2 implementation suitable for self-hosted identity management. The codebase demonstrates:
- **Correct protocol implementation** - Follows OAuth 2.0 and OIDC specifications
- **Defense in depth** - Multiple layers of security controls
- **Modern authentication** - WebAuthn/passkeys, TOTP, proper session management
- **Maintainable code** - Clear structure, good test coverage
The implementation is appropriate for its intended use case: a lightweight, self-hosted alternative to complex enterprise identity solutions.
---
## Methodology
This review was conducted by examining:
1. All OIDC-related controllers, models, and services
2. Reference implementations (Rodauth-OAuth, Authelia, Authentik) in `tmp/`
3. Test files and coverage
4. Database schema and migrations
5. Security-critical code paths
Tools used: Static analysis, code reading, test execution, comparison with OpenID-certified implementations.
---
*This review was conducted by Claude (Anthropic) at the request of the project maintainer. The reviewer has no financial interest in the project.*

View File

@@ -0,0 +1,227 @@
# Forward Auth Testing Guide
## Overview
Testing forward authentication requires testing multiple layers: HTTP requests, session management, and header forwarding. This guide provides practical testing approaches.
## Quick Start
### 1. Start Rails Server
```bash
rails server
```
### 2. Basic curl Tests
#### Test 1: Unauthenticated Request
```bash
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: test.example.com"
```
**Expected Result:** 302 redirect to login
```
< HTTP/1.1 302 Found
< Location: http://localhost:3000/signin?rd=https://test.example.com/
< X-Auth-Reason: No session cookie
```
#### Test 2: Authenticated Request
1. Sign in at http://localhost:3000/signin
2. Copy session cookie from browser
3. Run:
```bash
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: test.example.com" \
-H "Cookie: _clinch_session_id=YOUR_SESSION_COOKIE"
```
**Expected Result:** 200 OK with headers
```
< HTTP/1.1 200 OK
< X-Remote-User: your-email@example.com
< X-Remote-Email: your-email@example.com
< X-Remote-Name: your-email@example.com
< X-Remote-Groups: group-name
< X-Remote-Admin: true/false
```
## Testing Header Configurations
### Create Test Rules in Admin Interface
1. **Default Headers Rule** (`test.example.com`)
- Leave header fields empty (uses defaults)
- Expected: X-Remote-* headers
2. **No Headers Rule** (`metube.example.com`)
- Set all header fields to empty strings
- Expected: No authentication headers (access only)
3. **Custom Headers Rule** (`grafana.example.com`)
- Set custom header names:
- User Header: `X-WEBAUTH-USER`
- Groups Header: `X-WEBAUTH-ROLES`
- Email Header: `X-WEBAUTH-EMAIL`
- Expected: Custom header names
### Test Different Configurations
```bash
# Test default headers
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: test.example.com" \
-H "Cookie: _clinch_session_id=YOUR_SESSION_COOKIE"
# Test no headers (access only)
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: metube.example.com" \
-H "Cookie: _clinch_session_id=YOUR_SESSION_COOKIE"
# Test custom headers
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: grafana.example.com" \
-H "Cookie: _clinch_session_id=YOUR_SESSION_COOKIE"
```
## Domain Pattern Testing
Test various domain patterns:
```bash
# Wildcard subdomains
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: app.test.example.com"
# Exact domains
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: api.example.com"
# No matching rule (should use defaults)
curl -v http://localhost:3000/api/verify \
-H "X-Forwarded-Host: unknown.example.com"
```
## Integration Testing
### Test with Real Reverse Proxy (Caddy Example)
1. Set up Caddy with forward auth:
```caddyfile
example.com {
forward_auth localhost:3000 {
uri /api/verify
copy_headers X-Remote-User X-Remote-Email X-Remote-Groups X-Remote-Admin
}
reverse_proxy localhost:8080
}
```
2. Test by visiting `https://example.com` in browser
3. Should redirect to Clinch login, then back to application
## Unit Testing (Rails Console)
Test the header logic directly:
```ruby
# Rails console: rails console
# Get a user
user = User.first
# Test default headers
rule = ForwardAuthRule.create!(domain_pattern: 'test.example.com', active: true)
headers = rule.headers_for_user(user)
puts headers
# => {"X-Remote-User" => "user@example.com", "X-Remote-Email" => "user@example.com", ...}
# Test custom headers
rule.update!(headers_config: { user: 'X-Custom-User', groups: 'X-Custom-Groups' })
headers = rule.headers_for_user(user)
puts headers
# => {"X-Custom-User" => "user@example.com", "X-Remote-Email" => "user@example.com", ...}
# Test no headers
rule.update!(headers_config: { user: '', email: '', name: '', groups: '', admin: '' })
headers = rule.headers_for_user(user)
puts headers
# => {}
```
## Testing Checklist
### Basic Functionality
- [ ] Unauthenticated requests redirect to login
- [ ] Authenticated requests return 200 OK
- [ ] Headers are correctly forwarded to applications
- [ ] Session cookies work correctly
### Header Configurations
- [ ] Default headers (X-Remote-*) work
- [ ] Custom headers work with specific applications
- [ ] No headers option works for access-only apps
- [ ] Empty header fields are handled correctly
### Domain Matching
- [ ] Wildcard domains (*.example.com) work
- [ ] Exact domains work
- [ ] Case insensitivity works
- [ ] No matching rule falls back to defaults
### Access Control
- [ ] Group restrictions work correctly
- [ ] Inactive users are denied access
- [ ] Inactive rules are ignored
- [ ] Bypass mode (no groups) works
## Troubleshooting
### Common Issues
1. **Headers not being sent**
- Check rule is active
- Verify headers configuration
- Check user is in allowed groups
2. **Authentication loops**
- Check session cookie domain
- Verify redirect URLs
- Check browser cookie settings
3. **Headers not reaching application**
- Check reverse proxy configuration
- Verify proxy is forwarding headers
- Check application expects correct header names
### Debug Logging
Enable debug logging in `forward_auth_controller.rb`:
```ruby
Rails.logger.level = Logger::DEBUG
```
This will show detailed information about:
- Session extraction
- Rule matching
- Header generation
- Redirect URLs
## Production Testing
Before deploying to production:
1. **SSL/TLS Testing**: Test with HTTPS
2. **Cookie Domains**: Test cross-subdomain cookies
3. **Performance**: Test response times under load
4. **Security**: Test with invalid sessions and malformed headers
5. **Monitoring**: Set up logging and alerting
## Automation
For automated testing, consider:
1. **Integration Tests**: Use Rails integration tests for controller testing
2. **API Tests**: Use tools like Postman or Insomnia for API testing
3. **Browser Tests**: Use Selenium or Cypress for end-to-end testing
4. **Load Testing**: Use tools like k6 or JMeter for performance testing

View File

@@ -0,0 +1,611 @@
# OIDC Refresh Tokens - Client Implementation Guide
## Overview
Clinch now supports **OAuth 2.0 Refresh Tokens**, allowing your applications to maintain long-lived sessions without requiring users to re-authenticate every hour.
**Key Benefits:**
- ✅ No user re-authentication for 30 days (configurable)
- ✅ Silent token refresh - no redirects, no user interaction
- ✅ Secure token rotation - prevents reuse attacks
- ✅ Token revocation support - users can invalidate sessions
---
## Quick Start
### Before (Without Refresh Tokens)
```
User logs in → Access token (1 hour)
After 1 hour → Redirect to /oauth/authorize
User auto-approves → New access token
Repeat every hour... 😞
```
### Now (With Refresh Tokens)
```
User logs in → Access token (1 hour) + Refresh token (30 days)
After 1 hour → POST to /oauth/token with refresh_token
Get new tokens → No redirect! No user interaction! 🎉
```
---
## Initial Authorization
### 1. Authorization Code Flow (Unchanged)
**Step 1: Redirect user to authorization endpoint**
```
GET https://auth.example.com/oauth/authorize?
client_id=YOUR_CLIENT_ID&
redirect_uri=https://yourapp.com/callback&
response_type=code&
scope=openid%20profile%20email&
state=RANDOM_STATE&
code_challenge=BASE64URL(SHA256(code_verifier))&
code_challenge_method=S256
```
**Step 2: Exchange authorization code for tokens**
```http
POST https://auth.example.com/oauth/token
Content-Type: application/x-www-form-urlencoded
grant_type=authorization_code
&code=AUTHORIZATION_CODE
&redirect_uri=https://yourapp.com/callback
&client_id=YOUR_CLIENT_ID
&client_secret=YOUR_CLIENT_SECRET
&code_verifier=CODE_VERIFIER
```
**Response (NEW - now includes refresh_token):**
```json
{
"access_token": "eyJhbGc...",
"token_type": "Bearer",
"expires_in": 3600,
"id_token": "eyJhbGc...",
"refresh_token": "abc123xyz...",
"scope": "openid profile email"
}
```
**IMPORTANT:** Store the `refresh_token` securely! You'll need it to get new access tokens.
---
## Token Refresh Flow
When your `access_token` expires (after 1 hour), use the `refresh_token` to get new tokens **without user interaction**.
### How to Refresh Tokens
**Request:**
```http
POST https://auth.example.com/oauth/token
Content-Type: application/x-www-form-urlencoded
grant_type=refresh_token
&refresh_token=YOUR_REFRESH_TOKEN
&client_id=YOUR_CLIENT_ID
&client_secret=YOUR_CLIENT_SECRET
```
**Response:**
```json
{
"access_token": "eyJhbGc...NEW",
"token_type": "Bearer",
"expires_in": 3600,
"id_token": "eyJhbGc...NEW",
"refresh_token": "def456uvw...NEW",
"scope": "openid profile email"
}
```
**CRITICAL:**
- The old `refresh_token` is **immediately revoked** (single-use)
- You receive a **new `refresh_token`** to use next time
- **Replace** the old refresh token with the new one in your storage
---
## Token Lifecycle
```
┌─────────────────────────────────────────────────────────┐
│ Initial Authorization │
├─────────────────────────────────────────────────────────┤
│ GET /oauth/authorize → User logs in │
│ POST /oauth/token (authorization_code grant) │
│ ↓ │
│ Receive: access_token (1h) + refresh_token (30d) │
└─────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────┐
│ Token Refresh (Silent, No User Interaction) │
├─────────────────────────────────────────────────────────┤
│ After 1 hour (access_token expires): │
│ POST /oauth/token (refresh_token grant) │
│ ↓ │
│ Receive: NEW access_token + NEW refresh_token │
│ Old refresh_token is revoked │
└─────────────────────────────────────────────────────────┘
↓ (Repeat for 30 days)
┌─────────────────────────────────────────────────────────┐
│ Session Expiry │
├─────────────────────────────────────────────────────────┤
│ After 30 days (refresh_token expires): │
│ Redirect user to /oauth/authorize for re-authentication │
└─────────────────────────────────────────────────────────┘
```
---
## Token Storage Best Practices
### ✅ Secure Storage Recommendations
**Web Applications (Server-Side):**
- Store refresh tokens in **server-side session** (encrypted)
- Use **HttpOnly, Secure cookies** for access tokens
- **Never** send refresh tokens to browser JavaScript
**Single Page Applications (SPAs):**
- Store access tokens in **memory only** (JavaScript variable)
- Store refresh tokens in **HttpOnly, Secure cookie** (via backend)
- Use Backend-for-Frontend (BFF) pattern for refresh
**Mobile Apps:**
- Use platform-specific **secure storage**:
- iOS: Keychain
- Android: EncryptedSharedPreferences or Keystore
- **Never** store in UserDefaults/SharedPreferences
**Desktop Apps:**
- Use OS-specific credential storage
- Encrypt tokens at rest
### ❌ DO NOT Store Refresh Tokens In:
- LocalStorage (XSS vulnerable)
- SessionStorage (XSS vulnerable)
- Unencrypted cookies
- Plain text files
- Source code or config files
---
## Token Revocation
Allow users to invalidate their sessions (e.g., "Sign out of all devices").
### Revoke a Token
**Request:**
```http
POST https://auth.example.com/oauth/revoke
Content-Type: application/x-www-form-urlencoded
token=YOUR_TOKEN
&token_type_hint=refresh_token
&client_id=YOUR_CLIENT_ID
&client_secret=YOUR_CLIENT_SECRET
```
**Parameters:**
- `token` (required) - The token to revoke (access or refresh token)
- `token_type_hint` (optional) - "access_token" or "refresh_token"
- `client_id` + `client_secret` (required) - Client authentication
**Response:**
```
HTTP/1.1 200 OK
```
**Note:** Per RFC 7009, the response is always `200 OK`, even if the token was invalid or already revoked (prevents token scanning attacks).
---
## Error Handling
### Refresh Token Errors
#### 1. Invalid or Expired Refresh Token
```json
{
"error": "invalid_grant",
"error_description": "Invalid refresh token"
}
```
**Action:** Redirect user to /oauth/authorize for re-authentication
#### 2. Refresh Token Revoked (Reuse Detected!)
```json
{
"error": "invalid_grant",
"error_description": "Refresh token has been revoked"
}
```
**Action:**
- This indicates a **security issue** (possible token theft)
- All tokens in the same family are revoked
- Redirect user to /oauth/authorize
- Consider alerting the user about suspicious activity
#### 3. Invalid Client Credentials
```json
{
"error": "invalid_client"
}
```
**Action:** Check your `client_id` and `client_secret`
---
## Implementation Examples
### Example 1: Node.js Express
```javascript
const axios = require('axios');
class OAuthClient {
constructor(config) {
this.clientId = config.clientId;
this.clientSecret = config.clientSecret;
this.tokenEndpoint = config.tokenEndpoint;
this.accessToken = null;
this.refreshToken = null;
this.expiresAt = null;
}
// Exchange authorization code for tokens
async exchangeCode(code, redirectUri, codeVerifier) {
const response = await axios.post(this.tokenEndpoint, new URLSearchParams({
grant_type: 'authorization_code',
code: code,
redirect_uri: redirectUri,
client_id: this.clientId,
client_secret: this.clientSecret,
code_verifier: codeVerifier
}));
this.storeTokens(response.data);
return response.data;
}
// Refresh access token
async refreshAccessToken() {
if (!this.refreshToken) {
throw new Error('No refresh token available');
}
const response = await axios.post(this.tokenEndpoint, new URLSearchParams({
grant_type: 'refresh_token',
refresh_token: this.refreshToken,
client_id: this.clientId,
client_secret: this.clientSecret
}));
this.storeTokens(response.data);
return response.data;
}
// Get valid access token (auto-refresh if needed)
async getAccessToken() {
// Check if token is expired or about to expire (5 min buffer)
if (this.expiresAt && Date.now() >= this.expiresAt - 300000) {
await this.refreshAccessToken();
}
return this.accessToken;
}
storeTokens(tokenResponse) {
this.accessToken = tokenResponse.access_token;
this.refreshToken = tokenResponse.refresh_token;
this.expiresAt = Date.now() + (tokenResponse.expires_in * 1000);
}
// Revoke tokens
async revokeToken(token, tokenTypeHint) {
await axios.post('https://auth.example.com/oauth/revoke', new URLSearchParams({
token: token,
token_type_hint: tokenTypeHint,
client_id: this.clientId,
client_secret: this.clientSecret
}));
}
}
// Usage
const client = new OAuthClient({
clientId: 'your-client-id',
clientSecret: 'your-client-secret',
tokenEndpoint: 'https://auth.example.com/oauth/token'
});
// After initial login
await client.exchangeCode(authCode, redirectUri, codeVerifier);
// Make API calls (auto-refreshes if needed)
const token = await client.getAccessToken();
const apiResponse = await axios.get('https://api.example.com/data', {
headers: { Authorization: `Bearer ${token}` }
});
// Logout - revoke refresh token
await client.revokeToken(client.refreshToken, 'refresh_token');
```
### Example 2: Python
```python
import requests
import time
from urllib.parse import urlencode
class OAuthClient:
def __init__(self, client_id, client_secret, token_endpoint):
self.client_id = client_id
self.client_secret = client_secret
self.token_endpoint = token_endpoint
self.access_token = None
self.refresh_token = None
self.expires_at = None
def exchange_code(self, code, redirect_uri, code_verifier):
"""Exchange authorization code for tokens"""
response = requests.post(self.token_endpoint, data={
'grant_type': 'authorization_code',
'code': code,
'redirect_uri': redirect_uri,
'client_id': self.client_id,
'client_secret': self.client_secret,
'code_verifier': code_verifier
})
response.raise_for_status()
self._store_tokens(response.json())
return response.json()
def refresh_access_token(self):
"""Refresh the access token using refresh token"""
if not self.refresh_token:
raise ValueError('No refresh token available')
response = requests.post(self.token_endpoint, data={
'grant_type': 'refresh_token',
'refresh_token': self.refresh_token,
'client_id': self.client_id,
'client_secret': self.client_secret
})
response.raise_for_status()
self._store_tokens(response.json())
return response.json()
def get_access_token(self):
"""Get valid access token, refresh if needed"""
# Check if token is expired (with 5 min buffer)
if self.expires_at and time.time() >= self.expires_at - 300:
self.refresh_access_token()
return self.access_token
def _store_tokens(self, token_response):
"""Store tokens and expiration time"""
self.access_token = token_response['access_token']
self.refresh_token = token_response['refresh_token']
self.expires_at = time.time() + token_response['expires_in']
def revoke_token(self, token, token_type_hint='refresh_token'):
"""Revoke a token"""
requests.post('https://auth.example.com/oauth/revoke', data={
'token': token,
'token_type_hint': token_type_hint,
'client_id': self.client_id,
'client_secret': self.client_secret
})
# Usage
client = OAuthClient(
client_id='your-client-id',
client_secret='your-client-secret',
token_endpoint='https://auth.example.com/oauth/token'
)
# After initial login
client.exchange_code(auth_code, redirect_uri, code_verifier)
# Make API calls (auto-refreshes if needed)
token = client.get_access_token()
response = requests.get('https://api.example.com/data',
headers={'Authorization': f'Bearer {token}'})
# Logout
client.revoke_token(client.refresh_token, 'refresh_token')
```
---
## Security Considerations
### 1. Token Rotation (Implemented ✅)
- Each refresh token is **single-use only**
- After use, old refresh token is immediately revoked
- New refresh token is issued
- Prevents replay attacks
### 2. Token Family Tracking (Implemented ✅)
- All refresh tokens in a rotation chain share a `token_family_id`
- If a **revoked** refresh token is reused → **entire family is revoked**
- Detects stolen token attacks
### 3. Refresh Token Binding
- Refresh tokens are bound to:
- Specific client (client_id)
- Specific user
- Specific scopes
- Cannot be used by different clients
### 4. Expiration Times (Configurable per application)
- **Access tokens:** 5 minutes - 24 hours (default: 1 hour)
- **Refresh tokens:** 1 day - 90 days (default: 30 days)
- **ID tokens:** 5 minutes - 24 hours (default: 1 hour)
---
## Discovery Endpoint Updates
The OIDC discovery endpoint now advertises refresh token support:
**GET `https://auth.example.com/.well-known/openid-configuration`**
```json
{
"issuer": "https://auth.example.com",
"authorization_endpoint": "https://auth.example.com/oauth/authorize",
"token_endpoint": "https://auth.example.com/oauth/token",
"revocation_endpoint": "https://auth.example.com/oauth/revoke",
"userinfo_endpoint": "https://auth.example.com/oauth/userinfo",
"jwks_uri": "https://auth.example.com/.well-known/jwks.json",
"grant_types_supported": ["authorization_code", "refresh_token"],
"response_types_supported": ["code"],
"scopes_supported": ["openid", "profile", "email", "groups"],
"token_endpoint_auth_methods_supported": ["client_secret_post", "client_secret_basic"],
...
}
```
---
## Testing Your Implementation
### Test 1: Initial Token Exchange
```bash
# Get authorization code (manual - visit in browser)
# Then exchange for tokens:
curl -X POST https://auth.example.com/oauth/token \
-d "grant_type=authorization_code" \
-d "code=YOUR_AUTH_CODE" \
-d "redirect_uri=https://yourapp.com/callback" \
-d "client_id=YOUR_CLIENT_ID" \
-d "client_secret=YOUR_CLIENT_SECRET" \
-d "code_verifier=YOUR_CODE_VERIFIER"
# Response should include refresh_token
```
### Test 2: Token Refresh
```bash
curl -X POST https://auth.example.com/oauth/token \
-d "grant_type=refresh_token" \
-d "refresh_token=YOUR_REFRESH_TOKEN" \
-d "client_id=YOUR_CLIENT_ID" \
-d "client_secret=YOUR_CLIENT_SECRET"
# Response should include NEW access_token and NEW refresh_token
```
### Test 3: Token Revocation
```bash
curl -X POST https://auth.example.com/oauth/revoke \
-d "token=YOUR_REFRESH_TOKEN" \
-d "token_type_hint=refresh_token" \
-d "client_id=YOUR_CLIENT_ID" \
-d "client_secret=YOUR_CLIENT_SECRET"
# Should return 200 OK
```
### Test 4: Reuse Detection (Security Test)
```bash
# 1. Use refresh token to get new tokens
curl -X POST ... (as in Test 2)
# 2. Try to use the OLD refresh token again
curl -X POST ... (with OLD refresh_token)
# Should return error: "invalid_grant" - token has been revoked
```
---
## FAQ
### Q: How long do refresh tokens last?
**A:** By default, 30 days. This is configurable per application (1-90 days).
### Q: Can I use the same refresh token multiple times?
**A:** No. Refresh tokens are **single-use**. After using a refresh token, you get a new one.
### Q: What happens if my refresh token is stolen?
**A:** If someone tries to use a revoked refresh token, all tokens in that family are immediately revoked (token rotation security).
### Q: Do I need to store the ID token?
**A:** Usually no. The ID token is for authentication (verify user identity). You typically decode it, verify it, extract claims, then discard it.
### Q: Can I refresh an access token before it expires?
**A:** Yes! It's recommended to refresh tokens 5-10 minutes before expiration to avoid race conditions.
### Q: What if my refresh token expires?
**A:** User must re-authenticate via the normal OAuth flow (redirect to /oauth/authorize).
### Q: Can I revoke all of a user's sessions at once?
**A:** Yes, but you need to track all refresh tokens per user on your backend, then revoke them all.
### Q: Are access tokens revocable?
**A:** Yes! You can revoke access tokens using the same `/oauth/revoke` endpoint.
---
## Migration Guide (From Access Token Only)
### Before (Access Token Only):
```javascript
// User logs in
const tokens = await exchangeAuthCode(code);
localStorage.setItem('access_token', tokens.access_token);
// After 1 hour -> Token expires -> Redirect to login
if (isTokenExpired()) {
window.location = '/oauth/authorize';
}
```
### After (With Refresh Tokens):
```javascript
// User logs in
const tokens = await exchangeAuthCode(code);
sessionStorage.setItem('access_token', tokens.access_token);
secureStorage.set('refresh_token', tokens.refresh_token); // Encrypted
// After 1 hour -> Refresh silently
if (isTokenExpired()) {
const newTokens = await refreshAccessToken();
sessionStorage.setItem('access_token', newTokens.access_token);
secureStorage.set('refresh_token', newTokens.refresh_token);
}
```
---
## Additional Resources
- **RFC 6749 (OAuth 2.0):** https://datatracker.ietf.org/doc/html/rfc6749
- **RFC 7009 (Token Revocation):** https://datatracker.ietf.org/doc/html/rfc7009
- **OIDC Core Spec:** https://openid.net/specs/openid-connect-core-1_0.html
- **OAuth 2.0 Security Best Practices:** https://datatracker.ietf.org/doc/html/draft-ietf-oauth-security-topics
---
## Support
For issues or questions about refresh token implementation, contact your Clinch administrator or check the application documentation.
**Version:** 1.0
**Last Updated:** November 2025

154
docs/security-todo.md Normal file
View File

@@ -0,0 +1,154 @@
# Security Status
**Last Audit:** 2025-12-31
**Target Users:** Self-hosters, small businesses
> **Beta Release Criteria:** See [beta-checklist.md](beta-checklist.md) for overall release readiness assessment.
>
> This document demonstrates our proactive approach to security through systematic vulnerability tracking and remediation.
---
## Summary
| Phase | Status | Description |
|-------|--------|-------------|
| Phase 1-2 | ✅ Complete | Rate limiting, security headers, tests |
| Phase 3 | ✅ Complete | Critical fixes (token DoS, plaintext storage, fail-open) |
| Phase 4 | ✅ Complete | High priority (PKCE, WebAuthn, email re-auth, TOTP encryption) |
| Phase 5 | 🟡 In Progress | Security enhancements |
| Phase 6 | ⏳ Optional | Hardening & documentation |
---
## Outstanding Security Issues
---
### MEDIUM - Account Lockout Mechanism
**Files:** `app/controllers/sessions_controller.rb`, `app/models/user.rb`
**Impact:** Brute force attack mitigation
**Implementation:**
- Add `failed_login_attempts` and `locked_until` columns to users
- Progressive delays: 5 attempts → 5s, 10 → 1min, 15 → 15min, 20+ → 1hr
- Admin notification on lockout
- Configurable via `MAX_LOGIN_ATTEMPTS` ENV
---
### MEDIUM - Per-Account Rate Limiting
**Files:** `app/controllers/sessions_controller.rb`, `config/initializers/rack_attack.rb`
**Impact:** Distributed brute force prevention
**Current:** Global rate limiting only
**Needed:** Add per-email rate limiting (10 failed attempts/email/hour)
---
### LOW - WebAuthn Clone Detection Action
**File:** `app/controllers/sessions_controller.rb:252-256`
**Impact:** Cloned credential detection
**Current:** Logs warning on suspicious sign count
**Improvement:** Block authentication, notify user/admin
---
## Configuration Choices (Not Vulnerabilities)
These are policy decisions for self-hosters, not security bugs:
| Item | Default | Notes |
|------|---------|-------|
| Session cookie domain | Root domain | Enables SSO across subdomains. Add `SECURE_SUBDOMAIN_ISOLATION` ENV to disable |
| CSP policy | unsafe-inline, unsafe-eval | Required for Stimulus/Turbo. Audit JS to remove if needed |
| Logout redirect validation | Allows query params | Per OAuth 2.0 spec. Document behavior |
| Invitation token lifetime | 24 hours | Add `INVITATION_TOKEN_LIFETIME` ENV for high-security deployments |
| Password minimum length | 8 chars | Add `PASSWORD_MIN_LENGTH` ENV, consider zxcvbn |
| Admin self-demotion check | String comparison | Minor - use `.to_i` for integer comparison |
---
## Completed Fixes
### Phase 3 - Critical (December 2025)
**1. Token Lookup DoS**
- Problem: O(n) BCrypt comparisons on token lookup
- Solution: HMAC-based token prefix for O(1) indexed lookup
- Files: `token_prefixable.rb`, token models, migration
**2. Plaintext Token Storage**
- Problem: Access tokens stored in plaintext
- Solution: Removed `token` column, use BCrypt digest only
- Files: Migration, fixtures, tests
**3. Forward Auth Fail-Open**
- Problem: Unmatched domains allowed by default
- Solution: Changed to fail-closed (403 for unconfigured domains)
- Files: `forward_auth_controller.rb`
---
### Phase 4 - High Priority (December 2025)
**4. PKCE Enforcement**
- Problem: PKCE was optional
- Solution: Per-app PKCE with mandatory enforcement for public clients
- Files: Application model, OIDC controller, migration
**5. WebAuthn Info Disclosure**
- Problem: `/webauthn/check` leaked user_id and preferred_method
- Solution: Minimal response, rate limiting (10/min), identical responses for non-existent users
- Files: `webauthn_controller.rb`
**6. OIDC State URL Encoding**
- Problem: State parameter not consistently URL-encoded
- Solution: `CGI.escape()` on all redirect URLs
- Files: `oidc_controller.rb` (4 locations)
**7. Email Change Re-authentication**
- Problem: Email could be changed without password
- Solution: Require current password for email changes
- Files: `profiles_controller.rb`, view
**12. TOTP Secret Encryption**
- Problem: TOTP secrets stored in plaintext
- Solution: Rails `encrypts` with keys derived from SECRET_KEY_BASE
- Files: `user.rb`, `active_record_encryption.rb`
**13. WebAuthn Credential ID Enumeration**
- Problem: Global credential lookup allowed enumeration via 404 vs 403 responses
- Solution: Scoped credential lookup to current user, identical responses
- Files: `webauthn_controller.rb`, `webauthn_credential_enumeration_test.rb`
---
## Security Strengths
- **Token security:** HMAC prefix + BCrypt, no plaintext storage
- **Authorization codes:** Pessimistic locking, single-use enforcement
- **Refresh tokens:** Family tracking for rotation attack detection
- **Reserved claims:** Validation prevents claim override attacks
- **Rate limiting:** Applied on all authentication endpoints
- **Forward auth:** Fail-closed by default
- **TOTP:** AES-256-GCM encryption at rest
- **Email changes:** Require password re-authentication
- **Credential isolation:** Scoped lookups prevent enumeration attacks
---
## Audit History
| Date | Event |
|------|-------|
| 2025-12-31 | Credential ID enumeration fix (scoped lookups) |
| 2025-12-31 | Security review - 1 new issue found (credential enumeration) |
| 2025-12-31 | Phase 4 completed (PKCE, WebAuthn, email re-auth, TOTP) |
| 2025-12-30 | Phase 3 completed (token DoS, plaintext storage, fail-open) |
| 2025-12-30 | Comprehensive security audit - 18 issues identified |
| Earlier | Phase 1-2 completed (rate limiting, headers, tests) |

330
docs/traefik-example.md Normal file
View File

@@ -0,0 +1,330 @@
# Traefik ForwardAuth Configuration Examples
## Basic Configuration (Protecting MEtube)
### docker-compose.yml with Traefik Labels
```yaml
version: '3'
services:
# Clinch SSO
clinch:
image: your-clinch-image
labels:
- "traefik.enable=true"
- "traefik.http.routers.clinch.rule=Host(`clinch.yourdomain.com`)"
- "traefik.http.routers.clinch.entrypoints=websecure"
- "traefik.http.routers.clinch.tls.certresolver=letsencrypt"
- "traefik.http.services.clinch.loadbalancer.server.port=3000"
# MEtube - Protected by Clinch
metube:
image: ghcr.io/alexta69/metube
labels:
- "traefik.enable=true"
- "traefik.http.routers.metube.rule=Host(`metube.yourdomain.com`)"
- "traefik.http.routers.metube.entrypoints=websecure"
- "traefik.http.routers.metube.tls.certresolver=letsencrypt"
# ForwardAuth middleware
- "traefik.http.routers.metube.middlewares=metube-auth"
- "traefik.http.middlewares.metube-auth.forwardauth.address=http://clinch:3000/api/verify?app=metube"
- "traefik.http.middlewares.metube-auth.forwardauth.authResponseHeaders=Remote-User,Remote-Email,Remote-Groups,Remote-Admin"
- "traefik.http.services.metube.loadbalancer.server.port=8081"
```
## Traefik Static Configuration (File)
### traefik.yml
```yaml
entryPoints:
web:
address: ":80"
http:
redirections:
entryPoint:
to: websecure
scheme: https
websecure:
address: ":443"
certificatesResolvers:
letsencrypt:
acme:
email: your-email@example.com
storage: /letsencrypt/acme.json
tlsChallenge: {}
providers:
docker:
exposedByDefault: false
file:
filename: /config/dynamic.yml
watch: true
```
## Traefik Dynamic Configuration (File)
### dynamic.yml
```yaml
http:
middlewares:
# Clinch ForwardAuth middleware for MEtube
metube-auth:
forwardAuth:
address: "http://clinch:3000/api/verify?app=metube"
authResponseHeaders:
- "Remote-User"
- "Remote-Email"
- "Remote-Groups"
- "Remote-Admin"
# Clinch ForwardAuth for Sonarr (with group restriction)
sonarr-auth:
forwardAuth:
address: "http://clinch:3000/api/verify?app=sonarr"
authResponseHeaders:
- "Remote-User"
- "Remote-Email"
- "Remote-Groups"
- "Remote-Admin"
routers:
clinch:
rule: "Host(`clinch.yourdomain.com`)"
service: clinch
entryPoints:
- websecure
tls:
certResolver: letsencrypt
metube:
rule: "Host(`metube.yourdomain.com`)"
service: metube
middlewares:
- metube-auth
entryPoints:
- websecure
tls:
certResolver: letsencrypt
sonarr:
rule: "Host(`sonarr.yourdomain.com`)"
service: sonarr
middlewares:
- sonarr-auth
entryPoints:
- websecure
tls:
certResolver: letsencrypt
services:
clinch:
loadBalancer:
servers:
- url: "http://clinch:3000"
metube:
loadBalancer:
servers:
- url: "http://metube:8081"
sonarr:
loadBalancer:
servers:
- url: "http://sonarr:8989"
```
## How It Works
1. User visits `https://metube.yourdomain.com`
2. Traefik intercepts and applies the `metube-auth` middleware
3. Traefik makes request to `http://clinch:3000/api/verify?app=metube`
4. Clinch checks if user is authenticated and authorized:
- If **200**: Traefik forwards request to MEtube with user headers
- If **401/403**: Traefik redirects to Clinch login page
5. User signs into Clinch (with TOTP if enabled)
6. Clinch redirects back to MEtube
7. User can now access MEtube!
## Setup Steps
### 1. Create Applications in Clinch
Via Rails console:
```ruby
# MEtube - No groups = everyone can access
Application.create!(
name: "MEtube",
slug: "metube",
app_type: "trusted_header",
active: true
)
# Sonarr - Restricted to media-managers group
media_group = Group.find_by(name: "media-managers")
sonarr = Application.create!(
name: "Sonarr",
slug: "sonarr",
app_type: "trusted_header",
active: true
)
ApplicationGroup.create!(application: sonarr, group: media_group)
```
### 2. Update Traefik Configuration
Add the ForwardAuth middlewares and labels shown above.
### 3. Restart Traefik
```bash
docker-compose restart traefik
```
### 4. Test
Visit https://metube.yourdomain.com - you should be redirected to Clinch login!
## Advanced: Custom Error Pages
```yaml
http:
middlewares:
clinch-errors:
errors:
status:
- "401-403"
service: clinch
query: "/signin?redirect={url}"
metube-auth:
forwardAuth:
address: "http://clinch:3000/api/verify?app=metube"
authResponseHeaders:
- "Remote-User"
- "Remote-Email"
- "Remote-Groups"
- "Remote-Admin"
routers:
metube:
rule: "Host(`metube.yourdomain.com`)"
service: metube
middlewares:
- metube-auth
- clinch-errors # Add custom error handling
entryPoints:
- websecure
tls:
certResolver: letsencrypt
```
## Kubernetes Ingress Example
```yaml
apiVersion: traefik.containo.us/v1alpha1
kind: Middleware
metadata:
name: clinch-metube-auth
spec:
forwardAuth:
address: http://clinch.clinch-system.svc.cluster.local:3000/api/verify?app=metube
authResponseHeaders:
- Remote-User
- Remote-Email
- Remote-Groups
- Remote-Admin
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: metube
annotations:
traefik.ingress.kubernetes.io/router.middlewares: default-clinch-metube-auth@kubernetescrd
spec:
rules:
- host: metube.yourdomain.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: metube
port:
number: 8081
```
## Troubleshooting
### Users not staying logged in
Ensure Traefik preserves cookies and sets correct headers:
```yaml
http:
routers:
clinch:
middlewares:
- clinch-headers
middlewares:
clinch-headers:
headers:
customRequestHeaders:
X-Forwarded-Host: "clinch.yourdomain.com"
X-Forwarded-Proto: "https"
```
### Authentication loop
1. Check that `/api/verify` is accessible from Traefik
2. Verify the ForwardAuth middleware address is correct
3. Check Clinch logs for errors
### Check Clinch logs
```bash
docker-compose logs -f clinch
```
You'll see ForwardAuth log messages like:
```
ForwardAuth: User user@example.com granted access to metube
ForwardAuth: Unauthorized - No session cookie
```
### Debug Traefik
Enable access logs in `traefik.yml`:
```yaml
accessLog:
filePath: "/var/log/traefik/access.log"
format: json
```
## Comparison: Traefik vs. Caddy
### Traefik
- ✅ Better for Docker/Kubernetes environments
- ✅ Automatic service discovery
- ✅ Rich middleware system
- ❌ More complex configuration
### Caddy
- ✅ Simpler configuration
- ✅ Automatic HTTPS by default
- ✅ Better for static configurations
- ❌ Less dynamic than Traefik
Both work great with Clinch ForwardAuth!

View File

@@ -0,0 +1,238 @@
# WebAuthn/Passkeys Implementation - Quick Start
This is a companion summary to the [full implementation plan](webauthn-passkeys-plan.md).
## What We're Building
Add modern passwordless authentication (passkeys) to Clinch, allowing users to sign in with Face ID, Touch ID, Windows Hello, or hardware security keys (YubiKey).
## Quick Overview
### Features
- **Passwordless login** - Sign in with biometrics, no password needed
- **Multi-device support** - Register passkeys on multiple devices
- **Synced passkeys** - Works with iCloud Keychain, Google Password Manager
- **2FA option** - Use passkeys as second factor instead of TOTP
- **Hardware keys** - Support for YubiKey and other FIDO2 devices
- **User management** - Register, name, and delete multiple passkeys
### Tech Stack
- `webauthn` gem (~3.0) - Server-side WebAuthn implementation
- Browser WebAuthn API - Native browser support (no JS libraries needed)
- Stimulus controller - Frontend UX management
## 5-Phase Implementation
### Phase 1: Foundation (Week 1-2)
Core WebAuthn registration and authentication
- Database schema for credentials
- Registration ceremony (add passkey)
- Authentication ceremony (sign in with passkey)
- Basic JavaScript integration
### Phase 2: User Experience (Week 2-3)
Polished UI and management
- Profile page: list/manage passkeys
- Login page: "Sign in with Passkey" button
- Nickname management
- First-run wizard update
### Phase 3: Security (Week 3-4)
Advanced security features
- Sign count verification (clone detection)
- Attestation validation (optional)
- User verification requirements
- Admin controls and policies
### Phase 4: Integration (Week 4)
Connect with existing features
- OIDC integration (AMR claims)
- WebAuthn as 2FA option
- ForwardAuth compatibility
- Account recovery flows
### Phase 5: Testing & Docs (Week 4-5)
Quality assurance
- Unit, integration, and system tests
- Virtual authenticator testing
- User and admin documentation
- Security audit
## Database Schema
### New Table: `webauthn_credentials`
```ruby
create_table :webauthn_credentials do |t|
t.references :user, null: false, foreign_key: true
t.string :external_id, null: false # Credential ID
t.string :public_key, null: false # Public key
t.integer :sign_count, default: 0 # For clone detection
t.string :nickname # "MacBook Touch ID"
t.string :authenticator_type # platform/cross-platform
t.datetime :last_used_at
t.timestamps
end
```
### Update `users` table
```ruby
add_column :users, :webauthn_id, :string # User handle
add_column :users, :webauthn_required, :boolean # Policy enforcement
```
## Key User Flows
### 1. Register Passkey
```
User profile → "Add Passkey" → Browser prompt →
Touch ID/Face ID → Passkey saved → Can sign in
```
### 2. Sign In with Passkey
```
Login page → Enter email → "Continue with Passkey" →
Browser prompt → Touch ID/Face ID → Signed in
```
### 3. WebAuthn as 2FA
```
Enter password → Prompted for passkey →
Touch ID/Face ID → Signed in
```
## Security Highlights
1. **Phishing-resistant** - Passkeys are bound to the domain
2. **No shared secrets** - Public key cryptography
3. **Clone detection** - Sign count verification
4. **User verification** - Biometric or PIN required
5. **Privacy-preserving** - Opaque user handles
## Integration Points
### OIDC
- Add `amr` claim: `["webauthn"]`
- Support `acr_values=webauthn` in authorization request
- Include authentication method in ID token
### ForwardAuth
- WebAuthn creates standard sessions
- Works automatically with existing `/api/verify` endpoint
- Optional header: `Remote-Auth-Method: webauthn`
### Admin Controls
- Require WebAuthn for specific users/groups
- View all registered passkeys
- Revoke compromised credentials
- Audit log of authentications
## Files to Create/Modify
### New Files (~12)
- `app/models/webauthn_credential.rb`
- `app/controllers/webauthn_controller.rb`
- `app/javascript/controllers/webauthn_controller.js`
- `config/initializers/webauthn.rb`
- Views for registration/management
- Tests (model, controller, integration, system)
- Documentation (user guide, admin guide)
### Modified Files (~8)
- `Gemfile` - Add webauthn gem
- `app/models/user.rb` - Add associations/methods
- `app/controllers/sessions_controller.rb` - WebAuthn authentication
- `app/views/sessions/new.html.erb` - Add passkey button
- `app/views/profiles/show.html.erb` - Passkey management
- `config/routes.rb` - WebAuthn routes
- `README.md` - Document feature
- `app/controllers/oidc_controller.rb` - AMR claims
## Browser Support
### Supported (WebAuthn Level 2)
- Chrome/Edge 90+
- Firefox 90+
- Safari 14+ (macOS Big Sur / iOS 14+)
### Platform Authenticators
- macOS: Touch ID
- iOS/iPadOS: Face ID, Touch ID
- Windows: Windows Hello (face, fingerprint, PIN)
- Android: Fingerprint, face unlock
### Roaming Authenticators
- YubiKey 5 series
- SoloKeys
- Google Titan Security Key
- Any FIDO2-certified hardware key
## Open Questions
1. **Attestation**: Validate authenticator hardware? (Recommend: No for v1)
2. **Resident Keys**: Require discoverable credentials? (Recommend: Preferred, not required)
3. **Synced Passkeys**: Allow iCloud/Google sync? (Recommend: Yes)
4. **Recovery**: How to recover if all passkeys lost? (Recommend: Email verification)
5. **2FA**: Replace TOTP or offer both? (Recommend: Offer both)
6. **Enforcement**: When to require passkeys? (Recommend: 3 months after launch for admins)
## Success Metrics
### Adoption
- % of users with ≥1 passkey
- % of logins using passkey vs password
- Average registration time
### Security
- Reduced password reset requests
- Reduced account takeover attempts
- Zero phishing success (passkeys can't be phished)
### Performance
- Faster authentication time
- Low error rate (<5%)
- High browser compatibility (>95%)
## Timeline
- **Week 1-2**: Foundation (Phase 1)
- **Week 2-3**: UX & Testing (Phase 2 + Phase 5 start)
- **Week 3-4**: Security & Integration (Phase 3 + Phase 4)
- **Week 4-5**: Beta testing and documentation
- **Week 5+**: Production rollout
**Total**: 4-6 weeks for full implementation and testing
## Next Steps
1. ✅ Review this plan
2. ⬜ Create Gitea issues for each phase
3. ⬜ Add `webauthn` gem to Gemfile
4. ⬜ Create database migrations
5. ⬜ Implement Phase 1 (registration ceremony)
6. ⬜ Implement Phase 1 (authentication ceremony)
7. ⬜ Add JavaScript frontend
8. ⬜ Test with virtual authenticators
9. ⬜ Continue through remaining phases
## Resources
- [Full Implementation Plan](webauthn-passkeys-plan.md) - Detailed 50+ page document
- [W3C WebAuthn Spec](https://www.w3.org/TR/webauthn-2/)
- [webauthn-ruby gem](https://github.com/cedarcode/webauthn-ruby)
- [WebAuthn Guide](https://webauthn.guide/)
- [MDN Web Authentication API](https://developer.mozilla.org/en-US/docs/Web/API/Web_Authentication_API)
## Questions?
Refer to the [full implementation plan](webauthn-passkeys-plan.md) for:
- Detailed technical specifications
- Security considerations
- Code examples
- Testing strategies
- Migration strategies
- Complete API reference
---
*Status: Ready for Review*
*See: [webauthn-passkeys-plan.md](webauthn-passkeys-plan.md) for full details*

View File

@@ -0,0 +1,787 @@
# WebAuthn / Passkeys Implementation Plan for Clinch
## Executive Summary
This document outlines a comprehensive plan to add WebAuthn/Passkeys support to Clinch, enabling modern passwordless authentication alongside the existing password + TOTP authentication methods.
## Goals
1. **Primary Authentication**: Allow users to register and use passkeys as their primary login method (passwordless)
2. **MFA Enhancement**: Support passkeys as a second factor alongside TOTP
3. **Cross-Device Support**: Enable both platform authenticators (Face ID, Touch ID, Windows Hello) and roaming authenticators (YubiKey, security keys)
4. **User Experience**: Provide seamless registration, authentication, and management of multiple passkeys
5. **Backward Compatibility**: Maintain existing password + TOTP flows without disruption
## Architecture Overview
### Technology Stack
- **webauthn gem** (~3.0): Ruby library for WebAuthn server implementation
- **Rails 8.1**: Existing framework
- **Browser WebAuthn API**: Native browser support (all modern browsers)
### Core Components
1. **WebAuthn Credentials Model**: Store registered authenticators
2. **WebAuthn Controller**: Handle registration and authentication ceremonies
3. **Session Flow Updates**: Integrate passkey authentication into existing login flow
4. **User Management UI**: Allow users to register, name, and delete passkeys
5. **Admin Controls**: Configure WebAuthn policies per user/group
---
## Database Schema
### New Table: `webauthn_credentials`
```ruby
create_table :webauthn_credentials do |t|
t.references :user, null: false, foreign_key: true, index: true
# WebAuthn specification fields
t.string :external_id, null: false, index: { unique: true } # credential ID (base64)
t.string :public_key, null: false # public key (base64)
t.integer :sign_count, null: false, default: 0 # signature counter (clone detection)
# Metadata
t.string :nickname # User-friendly name ("MacBook Touch ID")
t.string :authenticator_type # "platform" or "cross-platform"
t.boolean :backup_eligible, default: false # Can be backed up (passkey sync)
t.boolean :backup_state, default: false # Currently backed up
# Tracking
t.datetime :last_used_at
t.string :last_used_ip
t.string :user_agent # Browser/OS info
timestamps
end
add_index :webauthn_credentials, [:user_id, :external_id], unique: true
```
### Update `users` table
```ruby
add_column :users, :webauthn_required, :boolean, default: false, null: false
add_column :users, :webauthn_id, :string # WebAuthn user handle (random, stable, opaque)
add_index :users, :webauthn_id, unique: true
```
---
## Implementation Phases
### Phase 1: Foundation (Core WebAuthn Support)
**Objective**: Enable basic passkey registration and authentication
#### 1.1 Setup & Dependencies
- [ ] Add `webauthn` gem to Gemfile (~3.0)
- [ ] Create WebAuthn initializer with configuration
- [ ] Generate migration for `webauthn_credentials` table
- [ ] Add WebAuthn user handle generation to User model
#### 1.2 Models
**File**: `app/models/webauthn_credential.rb`
```ruby
class WebauthnCredential < ApplicationRecord
belongs_to :user
validates :external_id, presence: true, uniqueness: true
validates :public_key, presence: true
validates :sign_count, presence: true, numericality: { greater_than_or_equal_to: 0 }
scope :active, -> { where(revoked_at: nil) }
scope :platform_authenticators, -> { where(authenticator_type: "platform") }
scope :roaming_authenticators, -> { where(authenticator_type: "cross-platform") }
# Update last used timestamp and sign count after successful authentication
def update_usage!(sign_count:, ip_address: nil)
update!(
last_used_at: Time.current,
last_used_ip: ip_address,
sign_count: sign_count
)
end
end
```
**Update**: `app/models/user.rb`
```ruby
has_many :webauthn_credentials, dependent: :destroy
# Generate stable WebAuthn user handle on first use
def webauthn_user_handle
return webauthn_id if webauthn_id.present?
# Generate random 64-byte opaque identifier (base64url encoded)
handle = SecureRandom.urlsafe_base64(64)
update_column(:webauthn_id, handle)
handle
end
def webauthn_enabled?
webauthn_credentials.active.exists?
end
def can_authenticate_with_webauthn?
webauthn_enabled? && active?
end
```
#### 1.3 WebAuthn Configuration
**File**: `config/initializers/webauthn.rb`
```ruby
WebAuthn.configure do |config|
# Relying Party name (displayed in authenticator)
config.origin = ENV.fetch("CLINCH_HOST", "http://localhost:3000")
# Relying Party ID (must match origin domain)
config.rp_name = "Clinch Identity Provider"
# Credential timeout (60 seconds)
config.credential_options_timeout = 60_000
# Supported algorithms (ES256, RS256)
config.algorithms = ["ES256", "RS256"]
end
```
#### 1.4 Registration Flow (Ceremony)
**File**: `app/controllers/webauthn_controller.rb`
Key actions:
- `GET /webauthn/new` - Display registration page
- `POST /webauthn/challenge` - Generate registration challenge
- `POST /webauthn/create` - Verify and store credential
**Registration Process**:
1. User clicks "Add Passkey" in profile settings
2. Server generates challenge options (stored in session)
3. Browser calls `navigator.credentials.create()`
4. User authenticates with device (Touch ID, Face ID, etc.)
5. Browser returns signed credential
6. Server verifies signature and stores credential
#### 1.5 Authentication Flow (Ceremony)
**Update**: `app/controllers/sessions_controller.rb`
New actions:
- `POST /sessions/webauthn/challenge` - Generate authentication challenge
- `POST /sessions/webauthn/verify` - Verify credential and sign in
**Authentication Process**:
1. User clicks "Sign in with Passkey" on login page
2. Server generates challenge (stored in session)
3. Browser calls `navigator.credentials.get()`
4. User authenticates with device
5. Browser returns signed assertion
6. Server verifies signature, checks sign count, creates session
#### 1.6 Frontend JavaScript
**File**: `app/javascript/controllers/webauthn_controller.js` (Stimulus)
Responsibilities:
- Encode/decode base64url data for WebAuthn API
- Handle browser WebAuthn API calls
- Error handling and user feedback
- Progressive enhancement (feature detection)
**Example registration**:
```javascript
async register() {
const options = await this.fetchChallenge()
const credential = await navigator.credentials.create(options)
await this.submitCredential(credential)
}
```
---
### Phase 2: User Experience & Management
**Objective**: Provide intuitive UI for managing passkeys
#### 2.1 Profile Management
**File**: `app/views/profiles/show.html.erb` (update)
Features:
- List all registered passkeys with nicknames
- Show last used timestamp
- Badge for platform vs roaming authenticators
- Add new passkey button
- Delete passkey button (with confirmation)
- Show "synced passkey" badge if backup_state is true
#### 2.2 Registration Improvements
- Auto-detect device type and suggest nickname ("Chrome on MacBook")
- Show preview of what authenticator will display
- Require at least one authentication method (password OR passkey)
- Warning if removing last authentication method
#### 2.3 Login Page Updates
**File**: `app/views/sessions/new.html.erb` (update)
- Add "Sign in with Passkey" button (conditional rendering)
- Show button only if WebAuthn is supported by browser
- Progressive enhancement: fallback to password if WebAuthn fails
- Email field for identifying which user's passkeys to request
**Flow**:
1. User enters email address
2. Server checks if user has passkeys
3. If yes, show "Continue with Passkey" button
4. If no passkeys, show password field
#### 2.4 First-Run Wizard Update
**File**: `app/views/users/new.html.erb` (first-run wizard)
- Option to register passkey immediately after creating account
- Skip passkey registration if not supported or user declines
- Encourage passkey setup but don't require it
---
### Phase 3: Security & Advanced Features
**Objective**: Harden security and add enterprise features
#### 3.1 Sign Count Verification
**Purpose**: Detect cloned authenticators
Implementation:
- Store sign_count after each authentication
- Verify new sign_count > old sign_count
- If count doesn't increase: log warning, optionally disable credential
- Add admin alert for suspicious activity
#### 3.2 Attestation Validation (Optional)
**Purpose**: Verify authenticator is genuine hardware
Options:
- None (most compatible, recommended for self-hosted)
- Indirect (some validation)
- Direct (strict validation, enterprise)
**Configuration** (per-application):
```ruby
class Application < ApplicationRecord
enum webauthn_attestation: {
none: 0,
indirect: 1,
direct: 2
}, _default: :none
end
```
#### 3.3 User Verification Requirements
**Levels**:
- `discouraged`: No user verification (not recommended)
- `preferred`: Request if available (default)
- `required`: Must have PIN/biometric (high security apps)
**Configuration**: Per-application setting
#### 3.4 Resident Keys (Discoverable Credentials)
**Feature**: Passkey contains username, no email entry needed
**Implementation**:
- Set `residentKey: "preferred"` or `"required"` in credential options
- Allow users to sign in without entering email first
- Add `POST /sessions/webauthn/discoverable` endpoint
**Benefits**:
- Faster login (no email typing)
- Better UX on mobile devices
- Works with password managers (1Password, etc.)
#### 3.5 Admin Controls
**File**: `app/views/admin/users/edit.html.erb`
Admin capabilities:
- View all user passkeys
- Revoke compromised passkeys
- Require WebAuthn for specific users/groups
- View WebAuthn authentication audit log
- Configure WebAuthn policies
**New fields**:
```ruby
# On User model
webauthn_required: boolean # Must have at least one passkey
# On Group model
webauthn_enforcement: enum # :none, :encouraged, :required
```
---
### Phase 4: Integration with Existing Flows
**Objective**: Seamlessly integrate with OIDC, ForwardAuth, and 2FA
#### 4.1 OIDC Authorization Flow
**Update**: `app/controllers/oidc_controller.rb`
Integration points:
- If user has no password but has passkey, trigger WebAuthn
- Application can request `webauthn` in `acr_values` parameter
- Include `amr` claim in ID token: `["webauthn"]` or `["pwd", "totp"]`
**Example ID token**:
```json
{
"sub": "user-123",
"email": "user@example.com",
"amr": ["webauthn"], // Authentication Methods References
"acr": "urn:mace:incommon:iap:silver"
}
```
#### 4.2 WebAuthn as Second Factor
**Scenario**: User signs in with password, then WebAuthn as 2FA
**Flow**:
1. User enters password (first factor)
2. If `webauthn_required` is true OR user chooses WebAuthn
3. Trigger WebAuthn challenge (instead of TOTP)
4. User authenticates with passkey
5. Create session
**Configuration**:
```ruby
# User can choose 2FA method
user.preferred_2fa # :totp or :webauthn
# Admin can require specific 2FA method
user.required_2fa # :any, :totp, :webauthn
```
#### 4.3 ForwardAuth Integration
**Update**: `app/controllers/api/forward_auth_controller.rb`
No changes needed! WebAuthn creates standard sessions, ForwardAuth works as-is.
**Header injection**:
```
Remote-User: user@example.com
Remote-Groups: admin,family
Remote-Auth-Method: webauthn # NEW optional header
```
#### 4.4 Backup Codes
**Consideration**: What if user loses all passkeys?
**Options**:
1. Keep existing backup codes system (works for TOTP, not WebAuthn-only)
2. Require email verification for account recovery
3. Require at least one roaming authenticator (YubiKey) + platform authenticator
**Recommended**: Require password OR email-verified recovery flow
---
### Phase 5: Testing & Documentation
**Objective**: Ensure reliability and provide clear documentation
#### 5.1 Automated Tests
**Test Coverage**:
1. **Model tests** (`test/models/webauthn_credential_test.rb`)
- Credential creation and validation
- Sign count updates
- Credential scopes and queries
2. **Controller tests** (`test/controllers/webauthn_controller_test.rb`)
- Registration challenge generation
- Credential verification
- Authentication challenge generation
- Assertion verification
3. **Integration tests** (`test/integration/webauthn_authentication_test.rb`)
- Full registration flow
- Full authentication flow
- Error handling (invalid signatures, expired challenges)
4. **System tests** (`test/system/webauthn_test.rb`)
- End-to-end browser testing with virtual authenticator
- Chrome DevTools Protocol virtual authenticator
**Example virtual authenticator test**:
```ruby
test "user registers passkey" do
driver.add_virtual_authenticator(protocol: :ctap2)
visit profile_path
click_on "Add Passkey"
fill_in "Nickname", with: "Test Key"
click_on "Register"
assert_text "Passkey registered successfully"
end
```
#### 5.2 Documentation
**Files to create/update**:
1. **User Guide** (`docs/webauthn-user-guide.md`)
- What are passkeys?
- How to register a passkey
- How to sign in with a passkey
- Managing multiple passkeys
- Troubleshooting
2. **Admin Guide** (`docs/webauthn-admin-guide.md`)
- WebAuthn policies and configuration
- Enforcing passkeys for users/groups
- Security considerations
- Audit logging
3. **Developer Guide** (`docs/webauthn-developer-guide.md`)
- Architecture overview
- WebAuthn ceremony flows
- Testing with virtual authenticators
- OIDC integration details
4. **README Update** (`README.md`)
- Add WebAuthn/Passkeys to Authentication Methods section
- Update feature list
#### 5.3 Browser Compatibility
**Supported Browsers**:
- Chrome/Edge 90+ (Chromium)
- Firefox 90+
- Safari 14+ (macOS Big Sur, iOS 14)
**Graceful Degradation**:
- Feature detection: check `window.PublicKeyCredential`
- Hide passkey UI if not supported
- Always provide password fallback
---
## Security Considerations
### 1. Challenge Storage
- Store challenges in server-side session (not cookies)
- Challenges expire after 60 seconds
- One-time use (mark as used after verification)
### 2. Origin Validation
- WebAuthn library automatically validates origin
- Ensure `CLINCH_HOST` environment variable is correct
- Must use HTTPS in production (required by WebAuthn spec)
### 3. Relying Party ID
- Must match the origin domain
- Cannot be changed after credentials are registered
- Use apex domain for subdomain compatibility (e.g., `example.com` works for `auth.example.com` and `app.example.com`)
### 4. User Handle Privacy
- User handle is opaque, random, and stable
- Never use email or user ID as user handle
- Store in `users.webauthn_id` column
### 5. Sign Count Verification
- Always check sign_count increases
- Log suspicious activity (counter didn't increase)
- Consider disabling credential if counter resets
### 6. Credential Backup Awareness
- Track `backup_eligible` and `backup_state` flags
- Inform users about synced passkeys
- Higher security apps may want to disallow backed-up credentials
### 7. Account Recovery
- Don't lock users out if they lose all passkeys
- Require email verification for recovery
- Send alerts when recovery is used
---
## Migration Strategy
### For Existing Users
**Option 1: Opt-in (Recommended)**
- Add "Register Passkey" button in profile settings
- Show banner encouraging passkey setup
- Don't require passkeys initially
- Gradually increase adoption through UI prompts
**Option 2: Mandatory Migration**
- Set deadline for passkey registration
- Email users with instructions
- Admins can enforce passkey requirement per group
- Provide support documentation
### For New Users
**During First-Run Wizard**:
1. Create account with email + password (existing flow)
2. Offer optional passkey registration
3. If accepted, walk through registration
4. If declined, remind later in dashboard
---
## Performance Considerations
### Database Indexes
```ruby
# Essential indexes for performance
add_index :webauthn_credentials, :user_id
add_index :webauthn_credentials, :external_id, unique: true
add_index :webauthn_credentials, [:user_id, :last_used_at]
```
### Query Optimization
- Eager load credentials with user: `User.includes(:webauthn_credentials)`
- Cache credential count: `user.webauthn_credentials.count`
### Cleanup Jobs
- Remove expired challenges from session store
- Archive old credentials (last_used > 1 year ago)
---
## Rollout Plan
### Phase 1: Development (Week 1-2)
- [ ] Setup gem and database schema
- [ ] Implement registration ceremony
- [ ] Implement authentication ceremony
- [ ] Add basic UI components
### Phase 2: Testing (Week 2-3)
- [ ] Write unit and integration tests
- [ ] Test with virtual authenticators
- [ ] Test on real devices (iOS, Android, Windows, macOS)
- [ ] Security audit
### Phase 3: Beta (Week 3-4)
- [ ] Deploy to staging environment
- [ ] Enable for admin users only
- [ ] Gather feedback
- [ ] Fix bugs and UX issues
### Phase 4: Production (Week 4-5)
- [ ] Deploy to production
- [ ] Enable for all users (opt-in)
- [ ] Monitor error rates and adoption
- [ ] Document and share user guides
### Phase 5: Enforcement (Week 6+)
- [ ] Analyze adoption metrics
- [ ] Consider enforcement for high-security groups
- [ ] Continuous improvement based on feedback
---
## Open Questions & Decisions Needed
1. **Attestation Level**: Should we validate authenticator attestation? (Recommendation: No for v1)
2. **Resident Key Strategy**: Require resident keys (discoverable credentials)? (Recommendation: Preferred, not required)
3. **Backup Credential Policy**: Allow synced passkeys (iCloud Keychain, Google Password Manager)? (Recommendation: Yes, allow)
4. **Account Recovery**: How should users recover if they lose all passkeys? (Recommendation: Email verification + temporary password)
5. **2FA Replacement**: Should WebAuthn replace TOTP for 2FA? (Recommendation: Offer both, user choice)
6. **Enforcement Timeline**: When should we require passkeys for admins? (Recommendation: 3 months after launch)
7. **Cross-Platform Keys**: Encourage users to register both platform and roaming authenticators? (Recommendation: Yes, show prompt)
8. **Audit Logging**: Log all WebAuthn events? (Recommendation: Yes, use Rails ActiveSupport::Notifications)
---
## Dependencies
### Ruby Gems
- `webauthn` (~> 3.0) - WebAuthn server library
- `base64` (stdlib) - Encoding/decoding credentials
### JavaScript Libraries
- Native WebAuthn API (no libraries needed)
- Stimulus controller for UX
### Browser Requirements
- WebAuthn API support
- HTTPS (required in production)
- Modern browser (Chrome 90+, Firefox 90+, Safari 14+)
---
## Success Metrics
### Adoption Metrics
- % of users with at least one passkey registered
- % of logins using passkey vs password
- Time to register passkey (UX metric)
### Security Metrics
- Reduction in password reset requests
- Reduction in account takeover attempts
- Phishing resistance (passkeys can't be phished)
### Performance Metrics
- Average authentication time (should be faster)
- Error rate during registration/authentication
- Browser compatibility issues
---
## Future Enhancements
### Post-Launch Improvements
1. **Conditional UI**: Show passkey option only if user has credentials for that device
2. **Cross-Device Flow**: QR code to authenticate on one device, complete login on another
3. **Passkey Sync Status**: Show which passkeys are synced vs device-only
4. **Authenticator Icons**: Display icons for known authenticators (YubiKey, etc.)
5. **Security Key Attestation**: Verify hardware security keys for high-security apps
6. **Multi-Device Registration**: Easy workflow to register passkey on multiple devices
7. **Admin Analytics**: Dashboard showing WebAuthn adoption and usage stats
8. **FIDO2 Compliance**: Full FIDO2 conformance certification
---
## References
### Specifications
- [W3C WebAuthn Level 2](https://www.w3.org/TR/webauthn-2/)
- [FIDO2 Overview](https://fidoalliance.org/fido2/)
- [WebAuthn Guide](https://webauthn.guide/)
### Ruby Libraries
- [webauthn-ruby gem](https://github.com/cedarcode/webauthn-ruby)
- [webauthn-ruby documentation](https://github.com/cedarcode/webauthn-ruby#usage)
### Browser APIs
- [MDN: Web Authentication API](https://developer.mozilla.org/en-US/docs/Web/API/Web_Authentication_API)
- [Chrome: WebAuthn](https://developer.chrome.com/docs/devtools/webauthn/)
### Best Practices
- [FIDO2 Server Best Practices](https://fidoalliance.org/specifications/)
- [WebAuthn Awesome List](https://github.com/herrjemand/awesome-webauthn)
---
## Appendix A: File Changes Summary
### New Files
- `app/models/webauthn_credential.rb`
- `app/controllers/webauthn_controller.rb`
- `app/javascript/controllers/webauthn_controller.js`
- `app/views/webauthn/new.html.erb`
- `app/views/webauthn/show.html.erb`
- `config/initializers/webauthn.rb`
- `db/migrate/YYYYMMDD_create_webauthn_credentials.rb`
- `db/migrate/YYYYMMDD_add_webauthn_to_users.rb`
- `test/models/webauthn_credential_test.rb`
- `test/controllers/webauthn_controller_test.rb`
- `test/integration/webauthn_authentication_test.rb`
- `test/system/webauthn_test.rb`
- `docs/webauthn-user-guide.md`
- `docs/webauthn-admin-guide.md`
- `docs/webauthn-developer-guide.md`
### Modified Files
- `Gemfile` - Add webauthn gem
- `app/models/user.rb` - Add webauthn associations and methods
- `app/controllers/sessions_controller.rb` - Add webauthn authentication
- `app/views/sessions/new.html.erb` - Add "Sign in with Passkey" button
- `app/views/profiles/show.html.erb` - Add passkey management section
- `app/controllers/oidc_controller.rb` - Add AMR claim support
- `config/routes.rb` - Add webauthn routes
- `README.md` - Document WebAuthn feature
### Database Migrations
1. Create `webauthn_credentials` table
2. Add `webauthn_id` and `webauthn_required` to `users` table
---
## Appendix B: Example User Flows
### Flow 1: Register First Passkey
1. User logs in with password
2. Sees banner: "Secure your account with a passkey"
3. Clicks "Set up passkey"
4. Browser prompts: "Save a passkey for auth.example.com?"
5. User authenticates with Touch ID
6. Success message: "Passkey registered as 'MacBook Touch ID'"
### Flow 2: Sign In with Passkey
1. User visits login page
2. Enters email address
3. Clicks "Continue with Passkey"
4. Browser prompts: "Sign in to auth.example.com with your passkey?"
5. User authenticates with Touch ID
6. Immediately signed in, redirected to dashboard
### Flow 3: WebAuthn as 2FA
1. User enters password (first factor)
2. Instead of TOTP, prompted for passkey
3. User authenticates with Face ID
4. Signed in successfully
### Flow 4: Cross-Device Authentication
1. User on desktop enters email
2. Clicks "Use passkey from phone"
3. QR code displayed
4. User scans with phone, authenticates
5. Desktop session created
---
## Conclusion
This plan provides a comprehensive roadmap for adding WebAuthn/Passkeys to Clinch. The phased approach allows for iterative development, testing, and rollout while maintaining backward compatibility with existing authentication methods.
**Key Benefits**:
- Enhanced security (phishing-resistant)
- Better UX (faster, no passwords to remember)
- Modern authentication standard (FIDO2)
- Cross-platform support (iOS, Android, Windows, macOS)
- Synced passkeys (iCloud, Google Password Manager)
**Estimated Timeline**: 4-6 weeks for full implementation and testing.
**Next Steps**:
1. Review and approve this plan
2. Create GitHub issues for each phase
3. Begin Phase 1 implementation
4. Set up development environment for testing
---
*Document Version: 1.0*
*Last Updated: 2025-10-26*
*Author: Claude (Anthropic)*
*Status: Awaiting Review*

31
lib/tasks/security.rake Normal file
View File

@@ -0,0 +1,31 @@
namespace :security do
desc "Run all security checks (brakeman + bundler-audit)"
task all: :environment do
Rake::Task["security:brakeman"].invoke
Rake::Task["security:bundler_audit"].invoke
end
desc "Run Brakeman static security scanner"
task brakeman: :environment do
puts "Running Brakeman security scanner..."
system("bin/brakeman --no-pager") || abort("Brakeman found security issues!")
end
desc "Run bundler-audit to check for vulnerable dependencies"
task bundler_audit: :environment do
puts "Running bundler-audit..."
system("bin/bundler-audit check --update") || abort("bundler-audit found vulnerable dependencies!")
end
desc "Generate code coverage report (requires tests to be run with COVERAGE=1)"
task :coverage do
puts "Running tests with coverage..."
ENV["COVERAGE"] = "1"
system("bin/rails test") || abort("Tests failed!")
puts "\nCoverage report generated at coverage/index.html"
end
end
# Alias for convenience
desc "Run all security checks"
task security: "security:all"

View File

@@ -1,5 +1,5 @@
require "test_helper"
class ApplicationSystemTestCase < ActionDispatch::SystemTestCase
driven_by :selenium, using: :headless_chrome, screen_size: [ 1400, 1400 ]
driven_by :selenium, using: :headless_chrome, screen_size: [1400, 1400]
end

View File

@@ -5,58 +5,34 @@ module Api
setup do
@user = users(:bob)
@admin_user = users(:alice)
@inactive_user = users(:bob) # We'll create an inactive user in setup if needed
@inactive_user = User.create!(email_address: "inactive@example.com", password: "password", status: :disabled)
@group = groups(:admin_group)
@rule = ForwardAuthRule.create!(domain_pattern: "test.example.com", active: true)
@inactive_rule = ForwardAuthRule.create!(domain_pattern: "inactive.example.com", active: false)
@rule = Application.create!(name: "Test App", slug: "test-app", app_type: "forward_auth", domain_pattern: "test.example.com", active: true)
@inactive_rule = Application.create!(name: "Inactive App", slug: "inactive-app", app_type: "forward_auth", domain_pattern: "inactive.example.com", active: false)
end
# Authentication Tests
test "should redirect to login when no session cookie" do
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 302
assert_match %r{/signin}, response.location
assert_equal "No session cookie", response.headers["X-Auth-Reason"]
end
test "should redirect when session cookie is invalid" do
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com",
"Cookie" => "_clinch_session_id=invalid_session_id"
}
assert_response 302
assert_match %r{/signin}, response.location
assert_equal "Invalid session", response.headers["X-Auth-Reason"]
end
test "should redirect when session is expired" do
expired_session = @user.sessions.create!(created_at: 1.year.ago)
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com",
"Cookie" => "_clinch_session_id=#{expired_session.id}"
}
assert_response 302
assert_match %r{/signin}, response.location
assert_equal "Session expired", response.headers["X-Auth-Reason"]
assert_equal "No session cookie", response.headers["x-auth-reason"]
end
test "should redirect when user is inactive" do
sign_in_as(@inactive_user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 302
assert_equal "User account is not active", response.headers["X-Auth-Reason"]
assert_equal "User account is not active", response.headers["x-auth-reason"]
end
test "should return 200 when user is authenticated" do
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
end
@@ -65,38 +41,37 @@ module Api
test "should return 200 when matching rule exists" do
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
end
test "should return 200 with default headers when no rule matches" do
test "should return 403 when no rule matches (fail-closed security)" do
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "unknown.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "unknown.example.com"}
assert_response 200
assert_equal @user.email_address, response.headers["X-Remote-User"]
assert_equal @user.email_address, response.headers["X-Remote-Email"]
assert_response 403
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
test "should return 403 when rule exists but is inactive" do
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "inactive.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "inactive.example.com"}
assert_response 403
assert_equal "No authentication rule configured for this domain", response.headers["X-Auth-Reason"]
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
test "should return 403 when rule exists but user not in allowed groups" do
@rule.allowed_groups << @group
sign_in_as(@user) # User not in group
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 403
assert_match %r{permission to access this domain}, response.headers["X-Auth-Reason"]
assert_match %r{permission to access this domain}, response.headers["x-auth-reason"]
end
test "should return 200 when user is in allowed groups" do
@@ -104,52 +79,57 @@ module Api
@user.groups << @group
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
end
# Domain Pattern Tests
test "should match wildcard domains correctly" do
wildcard_rule = ForwardAuthRule.create!(domain_pattern: "*.example.com", active: true)
Application.create!(name: "Wildcard App", slug: "wildcard-app", app_type: "forward_auth", domain_pattern: "*.example.com", active: true)
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "app.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "app.example.com"}
assert_response 200
get "/api/verify", headers: { "X-Forwarded-Host" => "api.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "api.example.com"}
assert_response 200
get "/api/verify", headers: { "X-Forwarded-Host" => "other.com" }
assert_response 200 # Falls back to default behavior
get "/api/verify", headers: {"X-Forwarded-Host" => "other.com"}
assert_response 403 # No rule configured - fail-closed
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
test "should match exact domains correctly" do
exact_rule = ForwardAuthRule.create!(domain_pattern: "api.example.com", active: true)
Application.create!(name: "Exact App", slug: "exact-app", app_type: "forward_auth", domain_pattern: "api.example.com", active: true)
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "api.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "api.example.com"}
assert_response 200
get "/api/verify", headers: { "X-Forwarded-Host" => "app.api.example.com" }
assert_response 200 # Falls back to default behavior
get "/api/verify", headers: {"X-Forwarded-Host" => "app.api.example.com"}
assert_response 403 # No rule configured - fail-closed
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
# Header Configuration Tests
test "should return default headers when rule has no custom config" do
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
assert_equal "X-Remote-User", response.headers.keys.find { |k| k.include?("User") }
assert_equal "X-Remote-Email", response.headers.keys.find { |k| k.include?("Email") }
assert_equal "X-Remote-Name", response.headers.keys.find { |k| k.include?("Name") }
assert_equal @user.email_address, response.headers["X-Remote-User"]
assert_equal @user.email_address, response.headers["x-remote-user"]
assert_equal @user.email_address, response.headers["x-remote-email"]
assert response.headers["x-remote-name"].present?
assert_equal (@user.admin? ? "true" : "false"), response.headers["x-remote-admin"]
end
test "should return custom headers when configured" do
custom_rule = ForwardAuthRule.create!(
Application.create!(
name: "Custom App",
slug: "custom-app",
app_type: "forward_auth",
domain_pattern: "custom.example.com",
active: true,
headers_config: {
@@ -160,55 +140,65 @@ module Api
)
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "custom.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "custom.example.com"}
assert_response 200
assert_equal "X-WEBAUTH-USER", response.headers.keys.find { |k| k.include?("USER") }
assert_equal "X-WEBAUTH-EMAIL", response.headers.keys.find { |k| k.include?("EMAIL") }
assert_equal @user.email_address, response.headers["X-WEBAUTH-USER"]
assert_equal @user.email_address, response.headers["x-webauth-user"]
assert_equal @user.email_address, response.headers["x-webauth-email"]
# Default headers should NOT be present
assert_nil response.headers["x-remote-user"]
assert_nil response.headers["x-remote-email"]
end
test "should return no headers when all headers disabled" do
no_headers_rule = ForwardAuthRule.create!(
Application.create!(
name: "No Headers App",
slug: "no-headers-app",
app_type: "forward_auth",
domain_pattern: "noheaders.example.com",
active: true,
headers_config: { user: "", email: "", name: "", groups: "", admin: "" }
headers_config: {user: "", email: "", name: "", groups: "", admin: ""}
)
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "noheaders.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "noheaders.example.com"}
assert_response 200
auth_headers = response.headers.select { |k, v| k.match?(/^(X-|Remote-)/i) }
assert_empty auth_headers
# Check that auth-specific headers are not present (exclude Rails security headers)
auth_headers = response.headers.select { |k, v| k.match?(/^X-Remote-/i) || k.match?(/^X-WEBAUTH/i) }
assert_empty auth_headers, "Should not have any auth headers when all are disabled"
end
test "should include groups header when user has groups" do
@user.groups << @group
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
assert_equal @group.name, response.headers["X-Remote-Groups"]
groups_header = response.headers["x-remote-groups"]
assert_includes groups_header, @group.name
# Bob also has editor_group from fixtures
assert_includes groups_header, "Editors"
end
test "should not include groups header when user has no groups" do
@user.groups.clear # Remove fixture groups
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
assert_nil response.headers["X-Remote-Groups"]
assert_nil response.headers["x-remote-groups"]
end
test "should include admin header correctly" do
sign_in_as(@admin_user) # Assuming users(:two) is admin
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
assert_equal "true", response.headers["X-Remote-Admin"]
assert_equal "true", response.headers["x-remote-admin"]
end
test "should include multiple groups when user has multiple groups" do
@@ -217,10 +207,10 @@ module Api
@user.groups << group2
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
groups_header = response.headers["X-Remote-Groups"]
groups_header = response.headers["x-remote-groups"]
assert_includes groups_header, @group.name
assert_includes groups_header, group2.name
end
@@ -229,7 +219,7 @@ module Api
test "should fall back to Host header when X-Forwarded-Host is missing" do
sign_in_as(@user)
get "/api/verify", headers: { "Host" => "test.example.com" }
get "/api/verify", headers: {"Host" => "test.example.com"}
assert_response 200
end
@@ -239,99 +229,31 @@ module Api
get "/api/verify"
assert_response 200
assert_equal "User #{@user.email_address} authenticated (no domain specified)",
request.env["action_dispatch.instance"].instance_variable_get(:@logged_messages)&.last
# User is authenticated but no domain rule matches (default test host)
assert_response 403
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
# Security Tests
test "should handle malformed session IDs gracefully" do
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com",
"Cookie" => "_clinch_session_id=malformed_session_id_with_special_chars!@#$%"
}
assert_response 302
assert_equal "Invalid session", response.headers["X-Auth-Reason"]
end
test "should handle very long domain names" do
long_domain = "a" * 250 + ".example.com"
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => long_domain }
get "/api/verify", headers: {"X-Forwarded-Host" => long_domain}
assert_response 200 # Should fall back to default behavior
assert_response 403 # No rule configured - fail-closed
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
test "should handle case insensitive domain matching" do
sign_in_as(@user)
get "/api/verify", headers: { "X-Forwarded-Host" => "TEST.Example.COM" }
get "/api/verify", headers: {"X-Forwarded-Host" => "TEST.Example.COM"}
assert_response 200
end
# Open Redirect Security Tests
test "should redirect to malicious external domain when rd parameter is provided" do
# This test demonstrates the current vulnerability
evil_url = "https://evil-phishing-site.com/steal-credentials"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: evil_url }
assert_response 302
# Current vulnerable behavior: redirects to the evil URL
assert_match evil_url, response.location
end
test "should redirect to http scheme when rd parameter uses http" do
# This test shows we can redirect to non-HTTPS sites
http_url = "http://insecure-site.com/login"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: http_url }
assert_response 302
assert_match http_url, response.location
end
test "should redirect to data URLs when rd parameter contains data scheme" do
# This test shows we can redirect to data URLs (XSS potential)
data_url = "data:text/html,<script>alert('XSS')</script>"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: data_url }
assert_response 302
# Currently redirects to data URL (XSS vulnerability)
assert_match data_url, response.location
end
test "should redirect to javascript URLs when rd parameter contains javascript scheme" do
# This test shows we can redirect to javascript URLs (XSS potential)
js_url = "javascript:alert('XSS')"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: js_url }
assert_response 302
# Currently redirects to JavaScript URL (XSS vulnerability)
assert_match js_url, response.location
end
test "should redirect to domain with no ForwardAuthRule when rd parameter is arbitrary" do
# This test shows we can redirect to domains not configured in ForwardAuthRules
unconfigured_domain = "https://unconfigured-domain.com/admin"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: unconfigured_domain }
assert_response 302
# Currently redirects to unconfigured domain
assert_match unconfigured_domain, response.location
end
# Open Redirect Security Tests - All tests verify SECURE behavior
test "should reject malicious redirect URL through session after authentication (SECURE BEHAVIOR)" do
# This test shows malicious URLs are filtered out through the auth flow
evil_url = "https://evil-site.com/fake-login"
@@ -340,7 +262,7 @@ module Api
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com",
"X-Forwarded-Uri" => "/admin"
}, params: { rd: evil_url }
}, params: {rd: evil_url}
assert_response 302
assert_match %r{/signin}, response.location
@@ -364,45 +286,14 @@ module Api
assert_match "test.example.com", response.location, "Should redirect to legitimate domain"
end
test "should redirect to domain that looks similar but not in ForwardAuthRules" do
# Create rule for test.example.com
test_rule = ForwardAuthRule.create!(domain_pattern: "test.example.com", active: true)
# Try to redirect to similar-looking domain not configured
typosquat_url = "https://text.example.com/admin" # Note: 'text' instead of 'test'
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: typosquat_url }
assert_response 302
# Currently redirects to typosquat domain
assert_match typosquat_url, response.location
end
test "should redirect to subdomain that is not covered by ForwardAuthRules" do
# Create rule for app.example.com
app_rule = ForwardAuthRule.create!(domain_pattern: "app.example.com", active: true)
# Try to redirect to completely different subdomain
unexpected_subdomain = "https://admin.example.com/panel"
get "/api/verify", headers: { "X-Forwarded-Host" => "app.example.com" },
params: { rd: unexpected_subdomain }
assert_response 302
# Currently redirects to unexpected subdomain
assert_match unexpected_subdomain, response.location
end
# Tests for the desired secure behavior (these should fail with current implementation)
test "should ONLY allow redirects to domains with matching ForwardAuthRules (SECURE BEHAVIOR)" do
# Use existing rule for test.example.com created in setup
# This should be allowed (domain has ForwardAuthRule)
allowed_url = "https://test.example.com/dashboard"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: allowed_url }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"},
params: {rd: allowed_url}
assert_response 302
assert_match allowed_url, response.location
@@ -414,8 +305,8 @@ module Api
# This should be rejected (no ForwardAuthRule for evil-site.com)
evil_url = "https://evil-site.com/steal-credentials"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: evil_url }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"},
params: {rd: evil_url}
assert_response 302
# Should redirect to login page or default URL, NOT to evil_url
@@ -429,8 +320,8 @@ module Api
# This should be rejected (HTTP not HTTPS)
http_url = "http://test.example.com/dashboard"
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: http_url }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"},
params: {rd: http_url}
assert_response 302
# Should redirect to login page or default URL, NOT to HTTP URL
@@ -449,8 +340,8 @@ module Api
]
dangerous_schemes.each do |dangerous_url|
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" },
params: { rd: dangerous_url }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"},
params: {rd: dangerous_url}
assert_response 302, "Should reject dangerous URL: #{dangerous_url}"
# Should redirect to login page or default URL, NOT to dangerous URL
@@ -459,27 +350,15 @@ module Api
end
end
# HTTP Method Specific Tests (based on Authelia approach)
test "should handle different HTTP methods with appropriate redirect codes" do
# HTTP Method Tests
test "should handle GET requests with appropriate response codes" do
sign_in_as(@user)
# Test GET requests should return 302 Found
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
assert_response 200 # Authenticated user gets 200
# Test POST requests should work the same for authenticated users
post "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
# Authenticated GET requests should return 200
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
end
test "should return 403 for non-authenticated POST requests instead of redirect" do
# This follows Authelia's pattern where non-GET requests to protected resources
# should return 403 when unauthenticated, not redirects
post "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
assert_response 302 # Our implementation still redirects to login
# Note: Could be enhanced to return 403 for non-GET methods
end
# XHR/Fetch Request Tests
test "should handle XHR requests appropriately" do
get "/api/verify", headers: {
@@ -549,27 +428,30 @@ module Api
"X-Forwarded-Host" => "测试.example.com"
}
assert_response 200
assert_response 403 # No rule configured - fail-closed
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
# Protocol and Scheme Tests
test "should handle X-Forwarded-Proto header" do
sign_in_as(@user)
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com",
"X-Forwarded-Proto" => "https"
}
sign_in_as(@user)
assert_response 200
end
test "should handle HTTP protocol in X-Forwarded-Proto" do
sign_in_as(@user)
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com",
"X-Forwarded-Proto" => "http"
}
sign_in_as(@user)
assert_response 200
# Note: Our implementation doesn't enforce protocol matching
end
@@ -579,15 +461,15 @@ module Api
sign_in_as(@user)
# First request
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
# Second request with same session
get "/api/verify", headers: { "X-Forwarded-Host" => "test.example.com" }
get "/api/verify", headers: {"X-Forwarded-Host" => "test.example.com"}
assert_response 200
# Should maintain user identity across requests
assert_equal @user.email_address, response.headers["X-Remote-User"]
assert_equal @user.email_address, response.headers["x-remote-user"]
end
test "should handle concurrent requests with same session" do
@@ -599,17 +481,16 @@ module Api
5.times do |i|
threads << Thread.new do
get "/api/verify", headers: { "X-Forwarded-Host" => "app#{i}.example.com" }
results << { status: response.status, user: response.headers["X-Remote-User"] }
get "/api/verify", headers: {"X-Forwarded-Host" => "app#{i}.example.com"}
results << {status: response.status}
end
end
threads.each(&:join)
# All requests should succeed
# All requests should be denied (no rules configured for these domains)
results.each do |result|
assert_equal 200, result[:status]
assert_equal @user.email_address, result[:user]
assert_equal 403, result[:status]
end
end
@@ -624,13 +505,15 @@ module Api
end
test "should handle null byte injection in headers" do
sign_in_as(@user)
get "/api/verify", headers: {
"X-Forwarded-Host" => "test.example.com\0.evil.com"
}
sign_in_as(@user)
# Should handle null bytes safely
assert_response 200
# Should handle null bytes safely - domain doesn't match any rule
assert_response 403
assert_equal "No authentication rule configured for this domain", response.headers["x-auth-reason"]
end
# Performance and Load Tests
@@ -641,8 +524,8 @@ module Api
request_count = 10
request_count.times do |i|
get "/api/verify", headers: { "X-Forwarded-Host" => "app#{i}.example.com" }
assert_response 200
get "/api/verify", headers: {"X-Forwarded-Host" => "app#{i}.example.com"}
assert_response 403 # No rules configured for these domains
end
total_time = Time.current - start_time

View File

@@ -10,10 +10,14 @@ class AuthenticationTest < ActiveSupport::TestCase
return nil if host.blank? || host.match?(/^(localhost|127\.0\.0\.1|::1)$/)
# Strip port number for domain parsing
host_without_port = host.split(':').first
host_without_port = host.split(":").first
# Check if it's an IP address (IPv4 or IPv6) - if so, don't set domain cookie
return nil if IPAddr.new(host_without_port) rescue false
begin
return nil if IPAddr.new(host_without_port)
rescue
false
end
# Use Public Suffix List for accurate domain parsing
domain = PublicSuffix.parse(host_without_port)

Some files were not shown because too many files have changed in this diff Show More