Phone Install
Build a release APK from the consumer repo and install it on a tethered Android device over wireless ADB. Use for phone sideload smoke tests, including consumer-specific build env overrides documented in the consumer repo. Accepts an ADB port and optional staging, development, production, APK-path, no-launch, or no-pull choices from the user's request.
Critique
PR-first adversarial review. Lean mode runs the highest-signal local Claude lenses; `deep` selects the full relevant matrix. Verified findings are posted inline before fixes, then replied to and resolved after a pushed fix.
Feature Dev
Guided feature development with codebase understanding and architecture focus
Issues
GitHub issue workflow — ready queue with dependency resolution, claim, start, close, link
Review Accessibility
Accessibility audit for a non-production web app containing only public or synthetic data — runs axe-core against approved routes, fixes violations confidently mapped to source, and opens a PR for human review. Works on any repo since nothing is installed in the target app; axe-core is injected at runtime through the browser tool.
Agent Loop
Autonomous issue implementation loop with strict issue allowlisting, one linked worktree and draft PR per issue, bounded local Codex/Claude review rounds, inline thread traceability, and fresh-base validation. Use for a bounded GitHub issue queue without hosted AI reviewers.
Aws Infra Migration
>-
Codex Review
Independent PR-first second opinion via the local Codex CLI. Codex finds issues read-only; Claude verifies them, posts confirmed findings inline before editing, then pushes, replies, and resolves. `verify` lets Codex run tests/build.
Diagnosing Bugs
Disciplined diagnosis loop for hard bugs and performance regressions. Use when something is broken, throwing, failing, flaky, or slow, or when the user says debug, diagnose, or "why is this happening".
Task Packet
Execute an implementation Task Packet end-to-end (code, tests, GitHub issue, PR, and closure)
Artifact
Publish a deliverable as a hosted Artifact page — choose page vs file vs comment, clear the synthetic-data gate, apply house tokens, publish or update in place. Use when work has an audience and a terminal dump or an unread markdown file would lose it — a prototype screen, a design walkthrough, a status report, a runbook, a data story. This is the workflow skill; Anthropic's artifact-design, arti…
Pr Stats
Pull PR-authorship counts per repo and per author across a GitHub org over a given time window. Use when the user asks "how many PRs did X open", "who's been shipping", "PR throughput by repo", or any similar curiosity about creation-volume.
Refactorpass
PR-first refactor pass that runs /simplify once against an open draft PR, verifies and commits the surviving cleanups, pushes, and records them in the PR ledger. Runs at most once per PR for this engine.
Reviewit
Post-push AI review orchestrator. Both modes fire Gemini Flash + Copilot at the same iteration watermark with staggered handling (Gemini first, Copilot folded in) — no in-skill /review. Lean caps at 2 iterations. `deep` arg bumps the cap to 4, early-exits when an iteration produced no fix-pushes, and runs a final /deepcritique so fresh agents look at the PR's current state.
Copilot Review
Address GitHub Copilot code review comments on a PR systematically
Grill
A relentless interview that stress-tests a plan or design until nothing is silently assumed.
Deepcritique
High-fidelity PR-first review chain that opens or reuses a draft PR, posts verified findings inline before fixes, and runs /critique deep — preceded by /refactorpass on this engine's first pass only. Rounds 3+ run in convergence mode. Use on complex or high-risk changes such as auth/crypto, schema migrations, sync-propagating work, and large refactors.
Actions Usage Audit
Audit an org's GitHub Actions minute consumption — build a month-over-month minutes-by-repo matrix from the billing usage API, attribute a spike to repos and workflows, and surface concrete cut opportunities. Use when an Actions/billing alert fires or you need to understand what is burning CI minutes. Read-only.
Backlog Refinement
Prepare the GitHub backlog for autonomous /agent-loop completion — assess each open issue against the agent-readiness rubric, auto-rewrite agent-shaped issues into agent-ready form (acceptance criteria, file pointers, out-of-scope guardrails) and tag dev:agent, exclude the rest with agent-bail:* reasons, and run the post-loop RCA aggregation that sharpens the rubric from every bail. Use before ru…