warden

Your code is under new management. Agents that review your code - locally or on every PR - using the Skills you already know and love.

Why Warden?

Skills, not prompts. Define analysis once, run it anywhere. Bootstrap your environment with skills from conventional directories (.agents/skills/ or .claude/skills/).

Two ways to run. CLI catches issues before you push. GitHub Action reviews every PR automatically.

GitHub-native. Eligible findings appear as inline PR comments with suggested fixes, and every finding is reported in Checks.

Quick Start

# Initialize warden in your repository
npx @sentry/warden init

# Add the built-in baseline reviews
npx @sentry/warden add security-review
npx @sentry/warden add code-review

# Run a pre-review on current branch changes
# Uses Pi. Set WARDEN_OPENAI_API_KEY, or WARDEN_ANTHROPIC_API_KEY for Anthropic models.
npx @sentry/warden

# Fix issues automatically
npx @sentry/warden --fix

Read the full documentation →

Contributing

git clone git@github.com:getsentry/warden.git
cd warden
pnpm install && pnpm build
pnpm test              # unit tests
pnpm test:coverage     # unit tests with LCOV coverage
pnpm evals             # end-to-end evals (requires API key)

See packages/evals/README.md for the eval framework.

License

FSL-1.1-ALv2

{
  "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
  "name": "warden",
  "description": "AI code review agents that watch over your code locally and on every PR",
  "owner": {
    "name": "dcramer",
    "email": "dcramer@gmail.com"
  },
  "plugins": [
    {
      "name": "warden",
      "description": "Run Warden and query Warden Service",
      "author": {
        "name": "Sentry"
      },
      "source": "./plugins/warden",
      "category": "code-review"
    }
  ]
}