SteveVitali/agent-skills

decompose-spec

Turn one large spec into an optimally-partitioned, dependency-ordered ticket plan — the fewest self-contained tickets that each fit one fresh-context implementation run — and seed a durable build ledger. The planning half of a multi-session build; pairs with…

原文语言：英语

orchestrate-build

Drive a decomposed spec to completion across fresh contexts: SETUP a dedicated worktree, run each ticket end-to-end through implement-spec in a fresh context, update the durable ledger, report progress, pause for intervention at ticket boundaries, and finish…

原文语言：英语

address-pr-comments

Work through unresolved review comments on your PR as the author: triage every thread, implement fixes where the reviewer is right, push back with rationale where they aren't, answer questions, commit + push, and reply to every thread with what happened. Use…

原文语言：英语

refresh-repo-docs

Audit and sync a repo's human-facing documentation (README, docs/, CHANGELOG, CONTRIBUTING, examples) against the actual code — detect drift deterministically, fix stale claims, remove cruft, fill gaps, and verify every claim written. Use when docs are…

原文语言：英语

review-pr

Review someone else's GitHub PR as a Staff Engineer and post a high-precision review — mechanical grounding (CI, local verification), focused design + security passes, calibrated severity labels, inline comments with suggestion blocks. Use when asked to…

原文语言：英语

agent-docs

Create or converge agent documentation (AGENTS.md hierarchy) for a codebase — bootstrap mode generates docs from scratch, refresh mode detects and fixes drift

原文语言：英语

implement-spec

Spec-driven, autonomous end-to-end: branch → plan + test matrix → implement (tests alongside) → self-review → gap-analysis vs spec → close gaps → rigorous downstream verification → PR + evidence report

原文语言：英语

self-review

Two-pass self-review of the current branch: mechanical verification (build/test/lint + checklists) then a Staff Engineer design critique

原文语言：英语