GITHUB AGENT SKILLS / 技能收录
给你的 AI 助手
装上专业技能
我们从 GitHub 持续收录、审核并分类优质的 Agent Skills，按职业与使用场景整理，复制一行命令即可安装到 Claude Code、Cursor 等智能体。
按职业浏览 Agent Skills
为不同岗位挑选的高频技能
按使用场景浏览 Agent Skills
从你要完成的任务出发
按创作者浏览 Agent Skills
按 GitHub Stars 排序的优秀创作者
Microsoft
@microsoft
openclaw
@openclaw
OpenAI
@openai
The Apache Software Foundation
@apache
GitHub
@github
Anthropic
@anthropics
AutoGPT
@Significant-Gravitas
Garry Tan
@garrytan全部 Skills 192262
writing-standards
Apply AS v2 technical-writing standards (clarity, brevity, active voice, Python/Django conventions). Use when writing a README, a docstring (PEP 257), a commit message, a PR description, an error message, or any technical doc.
verification-gate
Verify before claiming success. Use when about to say something works/is done, after implementing a feature, after a bugfix, or before pushing — run the proof command and confirm output first.
test-driven-development
Write the failing test before the implementation in AS v2 (pytest + Django/DRF). Use when implementing a feature or fixing a bug, before any production code is written.
subagent-development
Orquestra subagents no AS v2 — decompõe a tarefa em domínios independentes, paraleliza e revisa em 2 estágios. Use ao delegar exploração, implementação ou validação cross-arquivo para acelerar o trabalho e poupar o contexto principal.
security-scan
Automated security scanning for AS v2. Use before commits or when reviewing changes for vulnerabilities. Runs dependency audit (pip-audit, npm audit), secret detection (grep patterns), and code pattern scanning (SQL injection, XSS, eval, pickle, hardcoded secrets). Integrates with project hooks system.
resume-handoff
Resume work from a saved handoff document, validating its state against the current codebase before continuing. Use when resuming work from a saved handoff after a context clear or in a new session (path, session name, or GitHub issue).
receiving-code-review
Triage and respond to code review feedback for AS v2. Use when a PR has review comments, the user shares feedback, CI/linting flags issues, or after /review-staged or /review-enhanced runs.
carrier-relationship-management
Codified expertise for managing carrier portfolios, negotiating freight rates, tracking carrier performance, allocating freight, and maintaining strategic carrier relationships. Informed by transportation managers with 15+ years experience. Includes scorecarding frameworks, RFP processes, market intelligence, and compliance vetting. Use when managing carriers, negotiating rates, evaluating carrier performance, or building freight strategies.
ponytail
Modo "senior dev preguiçoso" contra over-engineering. Antes de escrever código, sobe uma escada e para no primeiro degrau que resolve: YAGNI → reusar o que já existe no codebase → stdlib → feature nativa da plataforma → dependência já instalada → uma linha → só então o mínimo que funciona. Deleção acima de adição, chato acima de esperto, o menor diff que funciona E que você entende. Use ao escrever código novo, adicionar feature, corrigir bug (causa-raiz, não sintoma) ou revisar um diff em busca de complexidade, abstrações ou dependências desnecessárias. Complementa (não substitui) as skills software-engineering e django-patterns. Snapshot fiel de DietrichGebert/ponytail (MIT).
plan-crosscheck
Adversarially cross-check an implementation plan against the REAL code before writing any of it. Fans out one verifier per plan-part that reads the code, confirms or refutes each claim with file:line, gives the concrete edit, and hunts for gaps — plus one skeptic that finds what the plan forgot. Use after drafting any non-trivial plan (multi-file, auth/RBAC/data-scope, migrations, many call-sites, contract/schema changes), BEFORE implementing.
performance-optimization
Profile and optimize AS v2 performance (Django backend + React frontend). Use when an endpoint is slow, a request fires N+1 queries, the bundle grows, or Core Web Vitals miss target.
memory-maintenance
Manter o dir de memórias (~/.claude/projects/.../memory/) e o índice MEMORY.md saudáveis — podar memória morta e corrigir o índice sem perder conhecimento. Use quando o dir inflar, o MEMORY.md encostar no cap de leitura (~24.4KB), ou você quiser limpar órfãs. Cobre a régua índice-vs-wikilink, a classificação A/B/C com grep -F, e o safety-grep obrigatório antes de deletar.
implement-plan
Implement an approved plan wave-by-wave with verification gates. Use when a plan path is given, when resuming after /clear, or when create-plan output is ready to build.
grill-with-docs
Entrevista implacável para afiar um plano ou design, registrando glossário e ADRs do Aprender Sistema durante a conversa.
graphify
Build/query a navigable knowledge graph of a codebase. Use when onboarding to an unfamiliar codebase, answering cross-module how-does-X-relate-to-Y, or tracing dependencies before a refactor.
git-worktrees
Work on multiple branches simultaneously without stashing or switching. Use when reviewing a PR while mid-feature, applying a hotfix without losing context, or comparing implementations side-by-side. Covers create/list/remove and AS v2 (Windows) conventions.
gate-calibration
Projetar um gate de CI que sobreviva — decidir se bloqueia, avisa ou usa ratchet; medir a precisão antes de escolher; limpar antes de trancar; e provar que ele não pode ser contornado. Use ao criar ou recalibrar qualquer check automático, especialmente sobre dívida preexistente.
frontend-ui-engineering
Build production-quality React UIs in AS v2 (React 19 + Antd v5 + Tailwind). Use when creating components, implementing pages, managing client state, or modifying any user-facing interface.