Skip to content

Repository files navigation

Warden

CI License: MIT

Open-source, AI-native QA platform. Warden reads a pull request's diff, selects the tests that matter, sends an AI agent to break the build, and posts a merge-gate verdict back to GitHub. Claude is the default engine, abstracted behind a provider interface so any model can be swapped in. Everything is self-hostable and MIT-licensed.

Documentation

Full docs live in docs/:

Packages

Package What it does
@warden/core Shared types, Zod schemas, and every platform interface
@warden/orchestrator Diff analysis, risk scoring, tier selection, and the merge gate
@warden/agent LLM providers (Claude default) + exploratory / generative / healer strategies
@warden/runner Playwright + Claude-Chrome browser engines and CTRF conversion
@warden/test-management SQLite execution history, YAML test cases, coverage, flake quarantine
@warden/reporter CTRF plus GitHub job-summary / PR-comment / check-run surfaces
@warden/cli The warden command-line tool
warden-action The published GitHub Action
@warden/design-system "Sentinel" — tokens, themes, and dashboard components
apps/dashboard The requirements-traceability dashboard

Also included: observability (Prometheus + Grafana), integrations (Linear / Jira / GitHub Projects), recorder (session recording → test synthesis), learning-studio, and dashboard-api.

Design system — "Sentinel"

A dark-first command center where test status is the palette (PASS / FAIL / FLAKY / BLOCKED / SKIPPED / QUARANTINED), the portcullis is the logo, and the quality gate is a live moment. Three themes ship — Signal (near-black, default), Watch (slate-teal), Day (light) — with a self-hosted Fira Code / Fira Sans pairing.

Develop

pnpm install
pnpm -w build       # build all packages
pnpm -w test        # run the full test suite
pnpm -w typecheck   # type-check all packages
pnpm -w lint        # formatting check

Requires Node 20+ and pnpm 10+.

License

MIT — see LICENSE.

About

Open-source, AI-native QA platform — reads a PR diff, selects tests, runs an AI agent to break the build, and posts a merge-gate verdict. Claude-first, self-hostable, MIT.

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

0 watching

Forks

Releases

Packages

Used by

Contributors

Languages