skills

Orchestrated multi-agent code review modeled on Cloudflare's system. Risk-tiers the diff, filters noise, fans out specialist reviewers (security, performance, code quality, docs, tests, release, AGENTS.md), and consolidates everything through a coordinator judge pass into one structured verdict. Use for "orchestrated review", "multi-agent code review", "deep review", "review like Cloudflare", "thorough AI code review", or to review a GitHub PR with severity-graded findings and an approve/block decision.

Document a recently solved problem into a searchable project-local knowledge store so each unit of engineering work makes the next one easier. Use after a non-trivial problem is solved and verified, when the user says "that worked" / "it's fixed" / "problem solved", or when invoked headless by the lfg pipeline. Writes one doc to docs/solutions/ and makes it discoverable.

Use when you have a written implementation plan to execute in a separate session with review checkpoints

Use when executing implementation plans with independent tasks in the current session

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

Use when you have a spec or requirements for a multi-step task, before touching code

Run the full autonomous engineering pipeline end-to-end (recall prior learnings, plan, implement with TDD, review and autofix, verify, commit, push, open PR, watch CI and fix failures, then compound the learning). Use ONLY when the user explicitly requests hands-off execution of a software task and provides a feature description. Do not auto-route casual conversation here.

Deep research on a technical problem using parallel subagents for web docs, Stack Overflow, and codebase exploration, saving a markdown report to docs/research/. Use when the user says "research X", "investigate X", "find out how X works", or needs evidence-based recommendations before implementing.

Whenever you create or substantially edit a GitHub Actions workflow (.github/workflows/*.yml), run `uvx zizmor --persona pedantic` on it and fix every finding before considering the file done. Use this any time you add a new workflow, paste a template workflow (e.g. the Claude GitHub App, a marketplace action, a generated CI/CD file), or change `uses:`, `permissions:`, or trigger blocks — even if the user didn't mention security or zizmor. A workflow that hasn't been run through the pedantic audit is not finished.

Exploratory browser QA — drive agent-browser through a real UI flow, capture screenshots and console output, then produce a structured markdown report with TC-## test cases, pass/fail verdicts, defects, risks and a ship recommendation. Use this whenever the user says "qa this", "verify like a qa engineer", "exploratory test", "manually verify in browser", "drive the browser and check X works", or asks for a QA report on a change. Prefer this over just running automated tests when the user explicitly wants hands-on, real-browser verification.

Generates a self-contained, deeply-technical "deep dive" blog post as a single offline-capable HTML file — AstroPaper aesthetic, light/dark toggle, Mermaid diagrams (CDN), Shiki syntax highlighting (CDN ESM). Use when the user types "$blog" or "$blog-post", asks to "write a blog post", "deep dive blog", "generate a blog post about how X works", "explain X in a blog post so I can rebuild it", "in-depth post about", or wants a long-form publishable HTML article reverse-engineering a codebase, library, or architecture (not a quick summary, not an interactive walkthrough).

Turn a topic into a structured blog-post draft for a developer audience, using The Good Docs Project templates as the structural backbone. Classifies the post intent (explainer, walkthrough, recipe, quickstart, debug story, release post, cheat sheet) via an internal index, fetches the matching Good Docs template (bundled locally from gitlab.com/tgdp/templates), then adapts it to blog voice (hook, narrative, takeaways, CTA). TRIGGER on: "good-docs-writer", "write a blog post about X", "draft a blog post on X", "blog post for X", or the explicit form "$good-docs-writer <intent> <topic>".

Vendor a library's source code into the current repo as a git subtree under `repos/<name>/`, then wire it into AGENTS.md as reference material so coding agents read real source instead of fighting stale docs. Use when the user types "$clone-repo <url>", "/clone-repo <url>", "vendor <url>", "clone <url> as a subtree", "add <url> as reference for the agent", "feed <url> to the agent", "just clone the repo <url>", or otherwise asks to set up a third-party repo as agent-readable context.

Generates annotated screenshot documentation for web apps and websites — local dev servers or live sites. Use when asked to "screenshot the app", "document the app", "visual docs", "screenshot documentation", "show me the app", "screenshot <url>", or "document this website".