Marshall-Hallenbeck/dot_files

complete-github-issue

Use when asked to work on, implement, or complete a GitHub issue — before writing any code, checks issue state, existing PRs, git history, and codebase to determine what work remains or if the issue is already resolved

原文语言：英语

claim-issue

Use when starting, checking, or finishing work on a GitHub issue in any repo where multiple agents may work in parallel — posts a session-info claim comment and manages the in-progress label so other agents skip claimed issues

原文语言：英语

create-pr

Use when asked to create, open, publish, or submit a pull request from the current branch

原文语言：英语

create-github-issue

Use when asked to create, open, file, or draft a GitHub issue

原文语言：英语

commit

Commit session changes only, or all changes if no session edits.

原文语言：英语

fix-tests

Autonomous test-fix pipeline. Runs full suite, diagnoses each failure, fixes source code, and iterates until all tests pass. Asks the user when uncertain about intended behavior.

原文语言：英语

safe-commit-all

Group ALL working-tree changes into logical commits.

原文语言：英语

claude-remote-control-service

Set up or repair a persistent Claude Remote Control environment for a project using a systemd user service, including blocker checks, project symlink, linger, verification, and URL capture.

原文语言：英语

sentry-cli

Guide for using the Sentry CLI to interact with Sentry from the command line. Use when the user asks about viewing issues, events, projects, organizations, making API calls, or authenticating with Sentry via CLI.

原文语言：英语

full-review

Holistic review pipeline: simplification, correctness, security, overcautious check, lint/fix, all tests, fix failures, audits, and final review. Fixes EVERYTHING — every severity and every warning — so nothing is left open.

原文语言：英语

run-quality-gate

Run full quality gate: linting, type checks, unit tests, and integration tests — all in parallel.

原文语言：英语

test-coverage-review

Review uncommitted changes for missing test coverage, then create the missing tests. Use after implementation and before commit.

原文语言：英语

to-the-point

Get straight to the point in every response. Use whenever responding to ANY user message — coding, debugging, explanations, planning, casual chat. Lead with the answer or next concrete action, number multi-step work, restate state across turns, suppress…

原文语言：英语

write-tests

Write comprehensive unit tests for a module or function. Handles test runner detection, mock conventions, and iterates until all tests pass.

原文语言：英语

summarize-changes

Summarize git changes for PRs or commits. Does not commit unless explicitly asked.

原文语言：英语

lint

Run formatting, linting, and type checking. Auto-detects project languages (Python, JS/TS, Bash). Fixes auto-fixable issues and stages changes.

原文语言：英语

overcautious-check

Use when reviewing changes before commit to catch patterns that silence failures, fail open, add unauthorized fallbacks, or hide errors behind graceful degradation.

原文语言：英语

security-review

Security audit of uncommitted changes. Scans for OWASP top 10, injection, exposed secrets, auth bypasses, and unsafe patterns.

原文语言：英语

optimize-setup

End-to-end Claude Code setup optimization: update CLAUDE.md with session learnings, recommend new automations, audit config quality, clean stale memory, and check for new capabilities.

原文语言：英语

run-team

Spawn an agent team in tmux split panes for parallel exploration. Best for investigation, debugging, ambiguous tasks, and multi-angle review where you want to steer work in real-time.

原文语言：英语

review

Review uncommitted code changes for bugs, security issues, and correctness. Finds AND fixes all issues.

原文语言：英语

run-integration-tests

Run integration-level suites. Auto-detects runner (pytest markers, Jest, E2E).

原文语言：英语

orchestrate-plan

Decompose a plan into parallel work tracks and execute with subagents. Reads PLAN.md, splits into non-overlapping tracks, spawns agents, and runs integration tests.

原文语言：英语

run-tests

Dispatcher skill for unit, integration, or full quality-gate workflows. Auto-detects language and scope.

原文语言：英语

run-unit-tests

Run unit-test suites. Auto-detects runner (pytest, Jest, Vitest, cargo, go test).

原文语言：英语

codex-review

Run Codex and Claude code reviews iteratively until both find no issues. Use after making changes and before committing.

原文语言：英语

lookup-docs

Look up documentation for any library using Context7. Resolves library IDs and queries versioned docs.

原文语言：英语

normalize-ai-config

Normalize agents/skills/hooks so responsibilities are separated and workflows are coherent.

原文语言：英语

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

原文语言：英语

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

原文语言：英语