Skills

Restate requirements, assess risks, create step-by-step implementation plan. Waits for confirmation before touching code.

Answer a quick side question without interrupting or losing context from the current task. Resumes work automatically.

Sequential and tmux/worktree orchestration guidance for multi-agent workflows.

Orchestrate parallel Claude Code agents via Claude DevFleet — plan from natural language, dispatch agents in isolated worktrees, monitor and read reports.

Multi-model collaborative development: Research → Ideation → Plan → Execute → Optimize → Review. Routes Frontend to Gemini, Backend to Codex.

Multi-model collaborative planning phase. Coordinates multiple AI models to produce a unified implementation plan.

Multi-model collaborative execution. Dispatches frontend and backend tasks to specialized models in parallel.

Frontend-focused multi-model development. Routes UI/UX tasks to the optimal model.

Backend-focused multi-model development. Routes API, DB, and server tasks to the optimal model.

General-purpose code review for quality, security, and maintainability. Flags critical, high, and medium issues.

PEP 8 compliance, Pythonic idioms, type hints, security, and performance review for Python projects.

Memory safety, modern C++ idioms (C++17/20), concurrency, and performance review based on C++ Core Guidelines.

Idiomatic Go patterns, concurrency safety, error handling, and interface design review.

Idiomatic Kotlin patterns, null safety, coroutine correctness, and Compose best practices review.

Ownership, lifetimes, error handling, unsafe usage, and idiomatic Rust patterns review.

Resolves TypeScript and build errors incrementally with minimal diffs. Focuses on getting the build green fast.

Fixes CMake, linker, and template errors in C++ projects with minimal changes.

Fixes go build, go vet warnings, and linter issues incrementally.

Fixes Kotlin/Gradle build errors, compiler warnings, and dependency issues.

Fixes cargo build errors, borrow checker issues, and Cargo.toml problems.

Resolves Gradle build errors for Android and KMP projects.

Generates and runs end-to-end tests with Playwright. Creates test journeys, quarantines flaky tests, uploads artifacts.

Analyzes test coverage and identifies uncovered paths. Reports gaps with actionable remediation suggestions.

TDD workflow for C++. Writes GoogleTest tests first, then implements. Validates with ctest.

Table-driven tests first, then implement. Validates with go test and race detector.

TDD with Kotest. Writes Kotest specs first, then implements. Validates with Gradle test task.

TDD for Rust. Writes unit and integration tests first, then implements. Validates with cargo test.

Extracts patterns from the session, self-evaluates quality before saving. Higher signal-to-noise than /learn alone.

Analyzes local git history to extract coding patterns and generate SKILL.md files. Local version of the Skill Creator GitHub App.

Shows skill portfolio health dashboard with charts, analytics, and maintenance suggestions.

Analyzes a draft prompt and outputs an optimized, ECC-enriched version ready to paste and run. Advisory only — does not execute.

Shows all learned instincts (project + global scope) with confidence scores.

Exports instincts from project or global scope to a shareable file.

Imports instincts from a file or URL into project or global scope.

Promotes project-scoped instincts to global scope so they apply across all projects.

Analyzes instinct usage patterns and suggests or generates evolved, consolidated structures.

Saves current session state to a dated file in ~/.claude/sessions/ for cross-session continuity.

Loads the most recent session file and resumes work with full context from where the last session ended.

Manages Claude Code session history, aliases, and metadata. Browse, tag, and retrieve past sessions.

Saves a working state checkpoint: git state, decisions made, and next steps. Resume from any checkpoint.

Lists all known projects with their instinct statistics and last-active timestamps.

Runs a full verification loop: type checks, tests, lints, and reports the overall health status of the project.

Enforces quality gate before merge: coverage, lint errors, type errors, and security checks must all pass.

Removes dead code, unused imports, and duplicate logic. Runs knip/depcheck/ts-prune to identify what is safe to delete.

Reads all project docs and cross-references with actual code to update stale documentation automatically.

Generates or updates codemap files that give Claude accurate knowledge of the project's file and module structure.

Analyzes context window usage across agents, skills, and MCP servers. Surfaces which items consume the most tokens.

Scans all active skills and extracts cross-cutting principles to distill into a unified rule file.

Converts a draft email into a preventive checklist — flags tone, missing context, implicit assumptions, and risk phrases.