pr-reviews
Review code changes on a given GitHub PR using gh CLI. Use when the user asks to review a pull request, analyze PR diffs, or provide feedback on open PRs with structured quality, security, and testing assessments.
outline-strong
Outline-Strong unified 5-layer validation - design all validation layers from requirements, then execute CREATE -> VERIFY -> INTEGRATE cycle. Use when implementing with comprehensive formal verification across types (Idris 2), specs (Quint), proofs (Lean 4), contracts, and tests simultaneously.
inits
Analyze a codebase and create an AGENTS.md file for future agent instances. Use when onboarding to a new repository, when the user asks to create or improve an AGENTS.md, or when documenting build/lint/test commands and high-level architecture.
hodd-rust
HODD-RUST validation-first Rust development - design Rust-specific verifications from requirements, then execute through validation pipeline. Use when developing Rust code with formal verification using rustfmt, clippy, static_assertions, Miri, Loom, Flux, contracts, Kani, or Lean4.
askme
Verbalized Sampling (VS) protocol for deep intent exploration before planning. Use when starting ambiguous or complex tasks, when multiple interpretations exist, or when you need to explore diverse intent hypotheses and ask maximum clarifying questions before committing to an approach.
ast-grep
Effective code search, analysis, and refactoring using ast-grep (sg). Use this skill for precise AST-based code modifications, structural search, and linting.
code-simplifier
Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality. Focuses on recently modified code unless instructed otherwise.
contexts
Coordinate context sweep before coding - gather relevant files, patterns, and tooling summaries. Use when preparing to implement a feature, fix, or refactor and need comprehensive architecture, pattern, tooling, and dependency context gathered first.
design-by-contract
Design-by-Contract (DbC) development - design contracts from requirements, then execute CREATE -> VERIFY -> TEST cycle. Use when implementing with formal preconditions, postconditions, and invariants using deal (Python), contracts (Rust), Zod (TypeScript), or Kotlin contracts.
gh-address-comments
Help address review/issue comments on the open GitHub PR for the current branch using gh CLI; verify gh auth first and prompt the user to authenticate if not logged in.
gh-fix-ci
Inspect GitHub PR checks with gh, pull failing GitHub Actions logs, summarize failure context, then create a fix plan and implement after user approval. Use when a user asks to debug or fix failing PR CI/CD checks on GitHub Actions and wants a plan + code changes; for external checks (e.g., Buildkite), only report the details URL and mark them out of scope.
resolve
Resolve code review comments by verifying their validity and proposing multiple solutions for confirmed issues. Use when addressing review feedback, analyzing whether review comments are valid, and generating architectural solutions (not naive fixes) for confirmed issues.
proof-driven
Proof-driven development with Lean 4 - design proofs from requirements, then execute CREATE -> VERIFY -> REMEDIATE cycle. Use when implementing with formal verification using Lean 4 theorems, lemmas, and proof tactics; zero-sorry policy enforced.
prompt-engineering
Interactive prompt optimization workflow for LLMs. Use when optimizing, improving, or engineering prompts for Claude, GPT, Gemini, or other language models; covers analysis, model-specific techniques, few-shot examples, XML structuring, and validation.
plans
Software architect and planning specialist - conduct thorough read-only planning before any action. Use when exploring a codebase to design implementation plans, defining objectives, gathering relevant files, and summarizing available tools before coding begins.
type-driven
Type-driven development with Idris 2 - design type specifications from requirements, then execute CREATE -> VERIFY -> IMPLEMENT cycle. Use when developing with dependent types, refined types, or proof-carrying types in Idris 2; totality and exhaustive pattern matching enforced.
validation-first
Validation-first development with Quint - design specifications from requirements, then execute CREATE -> VERIFY -> IMPLEMENT cycle. Use when developing with formal state machine specifications, invariants, and temporal properties using Quint before writing implementation code.
reviews
Review the code changes on the current branch. Use when the user asks to review their current work, analyze recent commits, or get a code quality assessment of the active branch against the main branch.
srgn-cli
Practical guide for building safe, syntax-aware srgn CLI commands for source-code search and transformation. Use when users ask for srgn commands, scoped refactors (comments/docstrings/imports/functions), multi-file rewrites with --glob, custom tree-sitter query usage, or CI-style checks with --fail-any/--fail-none.
test-driven
Test-Driven Development (TDD) - design tests from requirements, then execute RED -> GREEN -> REFACTOR cycle. Use when implementing features or fixes with TDD methodology, writing tests before code, or following XP-style development with pytest, vitest, cargo test, or go test.
keep
Session continuity for Keep workflow. Detects resumable work at conversation start.
ideation
Creative ideation for new project concepts. USE WHEN user says "new idea", "what if we built", "I want to create", "brainstorm a project", or discusses ideas without an existing project. Creates IDEA.md and project structure. Not for in-project exploration—use exploration skill for that.
exploration
Technical exploration within existing projects. USE WHEN user says "think through", "explore options", "investigate", "how should we approach", or needs to evaluate approaches before implementation. Creates exploration documents in project's explorations/ folder. Not for new project ideas—use ideation skill for that.
api-developer
API 개발 전문가. REST API, FastAPI, Flask, 인증, 문서화.
data-analyst
데이터 분석 전문가. pandas, numpy, 시각화, 통계 분석 지원.
korean-writer
한국어 작문 전문가. 문서 작성, 번역, 교정, 비즈니스 커뮤니케이션.
python-expert
Python 프로그래밍 전문가. 코드 작성, 디버깅, 최적화, 라이브러리 사용법 안내.
faf-expert
Expert in .faf (Foundational AI-context Format) files for persistent project context. Use when working with .faf files, project DNA, CLAUDE.md bi-sync, faf-cli commands, MCP server configuration, or AI-readiness scoring (0-100%). Updated for v2.8.0 Tool Visibility System.
ba
Task tracker for LLM sessions. Use "$ba ready" to see available work, "$ba claim <id>" to take ownership, "$ba finish <id>" when done.
optimize-claude-md
Analyze and optimize CLAUDE.md files for quality improvements and token reduction. Supports analysis-only, simplification-only, or full optimization modes. Use when asked to "review", "optimize", "simplify", "improve", "check", or "reduce tokens" in CLAUDE.md.
logic-validation
Validate logical consistency of decisions and claims through context forking. Use proactively when judgment validity is unclear, but prefer direct tools (Read, Bash) for simple verifications like file contents or test results.
execute-plan
Execute a pre-existing plan file by implementing tasks sequentially with test verification after each step. Use when plan/tasks are already decomposed and ready for implementation.
docs-audit
Comprehensive documentation audit and refresh. Checks code-documentation drift, generates prioritized report, then updates with approval. Use for: 'ドキュメント監査', 'docs audit', 'ドキュメント棚卸し', 'README全体見直し', 'documentation health check'.
decompose
Decompose complex tasks into detailed, actionable todos. Each todo has a rich description that is executable from the description alone.
commit-conventions
Use this skill when dealing with complex commits involving multiple change types, when unsure how to split changes into atomic commits, when determining appropriate scope for a commit, when writing commit body for non-trivial changes, or when the user asks about commit best practices, semantic versioning impact, or conventional commit format details.
codex-review-mcp
Use this skill when completing a plan in Plan mode, when stuck on a complex problem, when needing a second opinion on architecture decisions, when facing difficult debugging, or when the user explicitly asks for Codex review. Invokes OpenAI Codex MCP for external perspective and validation through conversational interface.
codex-review-cli
Use this skill when you need external perspective for code analysis, architecture advice, debugging guidance, or when stuck on complex problems. Executes Codex CLI in read-only mode for rapid consultation and second opinion. Note that Codex analyzes and suggests improvements but does not implement changes.
bugfix
Standardized bug fix workflow using TDD (Red-Green-Refactor). Use when fixing bugs, resolving issues, or when the user invokes /bugfix. Ensures root cause analysis, failing test first, then minimal fix.
clarify
Create detailed specifications by iteratively clarifying unclear points for Plan mode. Use when: After completing a plan when detailed requirements need clarification before implementation.
approach-check
Ask Claude to state its planned approach before writing code, for user approval. Use for medium-complexity tasks where Plan Mode is too heavy but wrong-approach risk exists.
verify-doc
Verify document self-containedness by having a context-free sub-agent explain its understanding. Use after writing specs, handoff docs, or any document that must be understood without conversation context.
validate-plan
Required when ExitPlanMode is blocked by hook. Reviews implementation plan for logical consistency using logic-validator. Adds <!-- validated --> marker after successful validation. Trigger keywords:"validate plan", "review plan", "check plan quality", "plan validation".
update-docs
Update documentation to match code changes. Triggered by: 'update docs', 'sync documentation', 'README outdated', 'docs更新', 'ドキュメント直して'. Delegates to subagent to protect context.
update-auto-approve
Analyze PermissionRequest decision logs and improve auto-approve configuration across all layers. Covers settings.json patterns, static rule engine, LLM evaluator prompt, and tests. Use when asked to "update auto-approve", "improve permissions", "reduce unnecessary asks", or "optimize permission rules".
task-handoff
Identify current session's task list and generate commands to start a new Claude session sharing the same tasks. Use when "handing off", "continuing later", "resuming tasks", or preparing work for next session.
task-enrich
Verify and enrich task descriptions to ensure each is independently executable with sufficient context. Use when "enriching tasks", "preparing handoff", "verifying task quality", or before running /task-handoff.
setup-claude-skills-for-web
Setup Claude Code skills from berlysia/dotfiles for web version in other projects. Creates .claude/settings.json with auto-update hook. Use when setting up skills for web version or when user asks to configure, install, or setup external Claude Code skills for web.
session-memo
Save current session's problem context as a memo file for cross-session reference. Uses sub-agent to minimize context consumption. Supports two modes - summarize current session or read a past session by ID.
self-correct-loop
Autonomous implement-build-test-fix loop with configurable max iterations. Shows results only after tests pass or iteration limit is reached. Use for feature implementation where tests already exist.
scope-guard
Analyze task scope and recommend execution strategy (sub-agent, multi-session, or scope cut). Orchestrates existing skills like /decompose, /task-enrich, and /task-handoff.
Page 1114 of 1643 · 82103 results
