Minimal but effective AI agent skill definitions in 100 lines or less.
rebuilt from skills.sh, favoring software, development, testing, and architecture workflows.
| no | name | src | desc | before | after |
|---|---|---|---|---|---|
| 1 | academic-paper-review | lingzhi227/agent-research-skills | review an academic paper with a structured review form and multiple reviewer personas | 74 | 31 |
| 2 | agent-memory | rohitg00/agentmemory | persist and recall context across sessions so the agent stops re-learning the same stack, bugs, and decisions | 43 | 33 |
| 3 | api-design | wshobson/agents | design a clear, consistent, and evolvable API contract before implementing it | 79 | 26 |
| 4 | architecture-decision-records | wshobson/agents | record a significant technical decision, its context, and consequences as an ADR | 302 | 27 |
| 5 | authentication | wshobson/agents | implement or review authentication flows, sessions, tokens, and password handling securely | 60 | 26 |
| 6 | authorization | wshobson/agents | design or review access control so users can only do what they are permitted to | 60 | 26 |
| 7 | bash | github/awesome-copilot | write or review robust shell scripts that fail fast and handle paths and input safely | 108 | 26 |
| 8 | brainstorming | obra/superpowers | widen the option space then converge on an ambiguous problem before committing to one approach | 114 | 25 |
| 9 | caching-strategies | bobmatnyc/claude-mpm-skills | add caching with the right layer, key design, invalidation, and consistency trade-offs | 107 | 26 |
| 10 | chrome-devtools | github/awesome-copilot | debug, inspect, automate, screenshot, or profile rendered browser behavior | 98 | 34 |
| 11 | ci-cd | wshobson/agents | design or improve a CI/CD pipeline for fast, reliable, automated build-test-deploy | 73 | 26 |
| 12 | clean-architecture | bobmatnyc/claude-mpm-skills | structure an app so business logic is independent of frameworks, UI, and databases | 137 | 25 |
| 13 | clean-codebase | bobmatnyc/claude-mpm-skills | clean up dead code, obsolete paths, and contradictions to keep a repo small and correct | 139 | 28 |
| 14 | concurrency | wshobson/agents | write or review concurrent code that avoids races, deadlocks, and shared-state bugs | 81 | 26 |
| 15 | cpp | github/awesome-copilot | write or review modern C++ with RAII, smart pointers, and value semantics | 235 | 26 |
| 16 | critique | pbakaus/impeccable | give structured, specific design critique with prioritized, actionable fixes | 126 | 27 |
| 17 | csharp | github/awesome-copilot | write or review modern, idiomatic C# with async, nullable reference types, and clean structure | 87 | 27 |
| 18 | database-migrations | wshobson/agents | write and apply schema migrations safely without downtime or data loss | 273 | 25 |
| 19 | dependency-injection | bobmatnyc/claude-mpm-skills | decouple components by injecting their dependencies instead of constructing them inline | 137 | 26 |
| 20 | design-patterns | bobmatnyc/claude-mpm-skills | apply a design pattern to a real problem without over-engineering the solution | 137 | 27 |
| 21 | dispatching-parallel-agents | obra/superpowers | split independent research, debugging, or coding tasks across isolated agents | 183 | 34 |
| 22 | docker-compose | github/awesome-copilot | define multi-service local and test environments with Docker Compose reliably | 605 | 26 |
| 23 | domain-driven-design | mattpocock/skills | model a complex domain with a shared language, bounded contexts, and clear aggregates | 50 | 26 |
| 24 | drizzle-orm | bobmatnyc/claude-mpm-skills | define type-safe Drizzle schemas, migrations, and queries with full TypeScript inference | 306 | 26 |
| 25 | duckdb-query | duckdb/duckdb-skills | run in-process analytical queries over CSV, Parquet, and JSON with DuckDB | 155 | 26 |
| 26 | error-handling | wshobson/agents | design error handling so failures are explicit, recoverable, and never silently swallowed | 94 | 26 |
| 27 | event-driven-architecture | wshobson/agents | design systems around events and messages with reliable, decoupled processing | 62 | 26 |
| 28 | executing-plans | obra/superpowers | execute written implementation plans with task checkpoints, verification, and blocker handling | 71 | 32 |
| 29 | extract-design-system | arvindrk/extract-design-system | recover tokens, components, and patterns from an existing codebase into a documented system | 46 | 26 |
| 30 | find-skills | vercel-labs/skills | find, compare, install, update, or remove agent skills from skills.sh or a source repo | 143 | 36 |
| 31 | finishing-a-development-branch | obra/superpowers | verify completed work and choose commit, PR, merge, leave-as-is, or cleanup flow | 252 | 31 |
| 32 | frontend-design | anthropics/skills | build or reshape UI with distinctive, intentional visual design instead of generic defaults | 30 | 26 |
| 33 | git-bisect | wshobson/agents | find the commit that introduced a bug with a binary search over history | 142 | 25 |
| 34 | git-commit | github/awesome-copilot | create conventional commits from the actual diff with logical staging and message analysis | 125 | 36 |
| 35 | git-rebase | wshobson/agents | clean up or reorder commits with rebase while keeping history safe and reviewable | 142 | 26 |
| 36 | github-actions | wshobson/agents | write or review GitHub Actions workflows that are fast, secure, and maintainable | 254 | 26 |
| 37 | go | github/awesome-copilot | write or review idiomatic Go with explicit errors, clear concurrency, and small interfaces | 287 | 26 |
| 38 | graphql | bobmatnyc/claude-mpm-skills | design or review GraphQL schemas and resolvers with safe, efficient data access | 1930 | 26 |
| 39 | grill-me | mattpocock/skills | pressure-test your own reasoning or plan, attacking the weakest assumptions first | 6 | 26 |
| 40 | grill-with-docs | mattpocock/skills | verify claims and API usage against fetched official docs instead of memory | 6 | 26 |
| 41 | handoff | mattpocock/skills | write a handoff so another session or agent resumes without context loss | 11 | 26 |
| 42 | improve-codebase-architecture | mattpocock/skills | propose high-leverage structural improvements and sequence them as safe steps | 42 | 26 |
| 43 | input-validation | bobmatnyc/claude-mpm-skills | validate and sanitize untrusted input at boundaries to prevent bad data and attacks | 685 | 26 |
| 44 | java | bobmatnyc/claude-mpm-skills | write or review modern, idiomatic Java with immutability, clear types, and clean structure | 787 | 27 |
| 45 | karpathy-guidelines | multica-ai/andrej-karpathy-skills | avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria | 47 | 47 |
| 46 | kotlin | github/awesome-copilot | write or review idiomatic Kotlin with null safety, immutability, and concise expressions | 412 | 26 |
| 47 | load-testing | bobmatnyc/claude-mpm-skills | measure system behavior under load to find limits, bottlenecks, and regressions | 107 | 26 |
| 48 | merge-conflict-resolution | mattpocock/skills | resolve git merge or rebase conflicts correctly without losing intended changes | 9 | 26 |
| 49 | mocking | bobmatnyc/claude-mpm-skills | use test doubles correctly — mock true boundaries, not the code under test | 147 | 26 |
| 50 | multi-stage-dockerfile | github/awesome-copilot | create or review optimized multi-stage Dockerfiles for smaller, safer application images | 47 | 32 |
| 51 | next-best-practices | vercel-labs/next-skills | build Next.js App Router apps with correct conventions, RSC boundaries, and data flow | 39 | 26 |
| 52 | next-cache-components | vercel-labs/next-skills | apply Next.js caching with use cache, cacheLife, cacheTag, PPR, and revalidation | 39 | 26 |
| 53 | nodejs | wshobson/agents | write or review Node.js backends with correct async, error handling, and resource limits | 35 | 26 |
| 54 | observability | wshobson/agents | instrument a system with logs, metrics, and traces to understand it in production | 55 | 26 |
| 55 | php | github/awesome-copilot | write or review modern, typed PHP with strict types, clear structure, and safe data handling | 102 | 26 |
| 56 | playwright-best-practices | currents-dev/playwright-best-practices-skill | write resilient Playwright tests with stable selectors, fixtures, parallelism, and CI | 261 | 26 |
| 57 | playwright-generate-test | github/awesome-copilot | generate Playwright tests from observed UI scenarios and iterate until they pass | 18 | 18 |
| 58 | postgresql-optimization | github/awesome-copilot | improve PostgreSQL schema, queries, indexes, and advanced Postgres feature usage | 405 | 33 |
| 59 | property-based-testing | bobmatnyc/claude-mpm-skills | test invariants over generated inputs instead of a few hand-picked examples | 325 | 26 |
| 60 | python | wshobson/agents | write or review idiomatic, typed, well-structured Python that is easy to test and maintain | 265 | 27 |
| 61 | rate-limiting | wshobson/agents | protect a service with rate limiting and backpressure without harming legitimate users | 79 | 26 |
| 62 | reactjs | bobmatnyc/claude-mpm-skills | write or review React components with correct hooks, state, effects, and rendering | 235 | 26 |
| 63 | receiving-code-review | obra/superpowers | handle review feedback by understanding, verifying, implementing, or pushing back with evidence | 214 | 32 |
| 64 | refactor | github/awesome-copilot | improve code structure, types, and maintainability while preserving external behavior | 646 | 34 |
| 65 | requesting-code-review | obra/superpowers | prepare completed work so a reviewer can evaluate and verify it quickly | 76 | 26 |
| 66 | research | lingzhi227/agent-research-skills | run a systematic literature review and synthesize sourced, well-reasoned findings | 182 | 30 |
| 67 | rest-api-design | bobmatnyc/claude-mpm-skills | design RESTful HTTP APIs with correct resources, methods, status codes, and pagination | 707 | 26 |
| 68 | ruby | github/awesome-copilot | write or review idiomatic, readable Ruby with clear objects and expressive methods | 110 | 26 |
| 69 | rust | github/awesome-copilot | write or review idiomatic Rust with sound ownership, error handling, and zero-cost abstractions | 108 | 27 |
| 70 | secrets-management | wshobson/agents | handle credentials, keys, and tokens without leaking them into code, logs, or history | 275 | 26 |
| 71 | security-review | github/awesome-copilot | review code or a design for security issues across the OWASP Top 10 and common pitfalls | 759 | 26 |
| 72 | shadcn | bobmatnyc/claude-mpm-skills | add, customize, and extend shadcn/ui components with correct Tailwind and theming | 1122 | 26 |
| 73 | skills-cli | xixu-me/skills | discover, install, list, check, update, remove, back up, restore, sync, or initialize Agent Skills with CLI commands | 282 | 39 |
| 74 | solid-principles | bobmatnyc/claude-mpm-skills | apply SOLID principles to make object-oriented code flexible and maintainable | 137 | 27 |
| 75 | sql | wshobson/agents | write or review correct, set-based SQL with safe parameters and sensible indexing | 165 | 26 |
| 76 | state-management | wshobson/agents | choose and structure client state so data stays consistent and predictable | 108 | 26 |
| 77 | structured-logging | bobmatnyc/claude-mpm-skills | emit structured, queryable logs with the right levels, context, and no sensitive data | 36 | 26 |
| 78 | subagent-driven-development | obra/superpowers | orchestrate specialized subagents across distinct parts of one larger task | 335 | 26 |
| 79 | swift | github/awesome-copilot | write or review idiomatic Swift with value types, optionals, and safe concurrency | 419 | 26 |
| 80 | systematic-debugging | obra/superpowers | find a bug's true cause with a hypothesis-and-test loop before editing | 220 | 25 |
| 81 | tailwind-design-system | wshobson/agents | implement a Tailwind design system with tokens, variants, and consistent components | 141 | 26 |
| 82 | teach | mattpocock/skills | explain code or a concept at the right depth for the asker, with real examples | 82 | 25 |
| 83 | test-driven-development | obra/superpowers | drive changes with a failing-test-first TDD loop, then refactor | 279 | 25 |
| 84 | threat-modeling | wshobson/agents | systematically identify threats to a system and decide how to mitigate them | 50 | 25 |
| 85 | to-issues | mattpocock/skills | break plans or PRDs into independently grabbable AFK/HITL vertical-slice issues | 84 | 34 |
| 86 | to-prd | mattpocock/skills | synthesize current conversation and repo context into a product requirements document | 77 | 34 |
| 87 | turborepo | bobmatnyc/claude-mpm-skills | configure a Turborepo monorepo with task pipelines, caching, and affected-only runs | 1498 | 26 |
| 88 | typescript | bobmatnyc/claude-mpm-skills | write or review TypeScript that maximizes type safety without fighting the compiler | 212 | 26 |
| 89 | typescript-advanced-types | wshobson/agents | model complex APIs with discriminated unions, conditional, and template literal types | 235 | 26 |
| 90 | understand-codebase | Lum1104/Understand-Anything | map an unfamiliar codebase's structure, dependencies, and business logic into a searchable knowledge graph | 258 | 33 |
| 91 | using-git-worktrees | obra/superpowers | isolate feature work in an existing or new git worktree without disrupting the current checkout | 216 | 34 |
| 92 | vercel-composition-patterns | vercel-labs/agent-skills | design composable React component APIs that scale without boolean prop sprawl | 65 | 26 |
| 93 | vercel-react-best-practices | vercel-labs/agent-skills | write or review React/Next.js for performance: waterfalls, re-renders, bundle size | 120 | 26 |
| 94 | verification-before-completion | obra/superpowers | require fresh evidence before claiming work is complete, fixed, passing, or ready | 140 | 38 |
| 95 | watch-video | bradautomates/claude-video | answer questions about a video URL or local file from extracted frames and a timestamped transcript | 123 | 31 |
| 96 | web-design-guidelines | vercel-labs/agent-skills | review frontend design, UX, accessibility, and visual polish against practical interface guidelines | 40 | 35 |
| 97 | webapp-testing | anthropics/skills | test a web app across unit, integration, and e2e at the cheapest honest layer | 74 | 25 |
| 98 | write-a-skill | mattpocock/skills | create compact agent skills with correct frontmatter, triggers, structure, and validation | 118 | 33 |
| 99 | writing-plans | obra/superpowers | write a structured, verifiable implementation plan before a complex task | 116 | 26 |
| 100 | zoom-out | mattpocock/skills | map unfamiliar code one abstraction up before choosing the next implementation action | 8 | 8 |