my-agents

Judge whether any iterative loop is well-designed — a verification loop, review loop, maker/checker loop, evaluator-optimizer, or agent self-improvement loop — or design a new one from scratch. Provides a 10-dimension quality rubric (Mode A) and a 9-step build guide (Mode B) grounded in TDD, Reflexion, LLM-as-judge, process-supervision, and agentic-workflow research. Use when asked "is this loop good enough", "review/critique this loop", "design a loop to catch X", "what should a verification loop cover", or whenever you are designing, modifying, or evaluating the mechanics of a produce → check → fix → re-check cycle — even if the word "loop" is never said. Do NOT use for executing an existing concrete loop (run that loop itself), one-shot scripts, or plain Q&A.

Audit a repository's control-plane health and produce an actionable improvement report. Use when the user asks to "audit the repo", "check repo health", "assess control plane", "entropy audit for the control system", "check what's missing for agents", or wants to know what's missing in their instruction files, rules, and verification infrastructure. Do NOT use for code review (use review), dependency audit (use dependency-audit), general documentation work (use project-documentation), or reviewing a specific PR or diff (use review or entropy-review).

Review code changes for entropy introduction: naming drift, error-model splits, dependency violations, doc staleness, and pattern duplication. Activate when the user asks for "entropy review", "consistency check", "drift check on this PR", "check if this change introduces new patterns", or wants a review focused on naming, error handling, or structural consistency rather than correctness. Do NOT activate for general code review (use review), full repo audit (use repo-entropy-audit), control-plane audit (use control-plane-auditor), brainstorming, or implementation tasks.

Full-repository entropy scan across six axes: structure (dependency graph, layering, SCC), semantics (naming diversity, state model fragmentation), behavior (error model variants, boundary handling), context (doc freshness, instruction-file coverage), protocol (instruction-file consistency), and control (rule enforcement gaps). Produces a module-level heatmap, trend comparison, and prioritized cleanup targets. Activate when the user asks for "entropy audit", "repo health scan", "heatmap", "tech debt scan", or "full entropy report". Do NOT use for PR-level review (use entropy-review) or control-plane audit (use control-plane-auditor).

REQUIRED for Claude Code or Codex CLI hook queries — has accurate cross-platform reference. Triggers on PreToolUse, PostToolUse, Stop, SessionStart hooks, .claude/settings.json or .codex/hooks.json config, prompt vs command hook types, hooks not firing, exit codes, stdin JSON, matchers, path blocking, or dual-platform hook setups. Skip for React hooks, webhooks, GitHub Actions.

Analyze a coding project and design a tailored AI development environment: recommend agents, skills, instruction files, MCP servers, and workflow composition across Claude Code and Codex. Also audit an existing AI setup for gaps, overlap, or bloat. Use when setting up, overhauling, or auditing a project's AI tooling as a whole, not for individual agent edits or skill management.

Manage the lifecycle of sub-agents: create or update agent packages, validate cross-surface agent contracts, evaluate routing behavior, tighten invocation boundaries, tune Codex runtime defaults, install or publish agents, or audit an agent library. Use only when the request is explicitly about agents or agent libraries, not for ordinary coding, implementation, or general project planning.

Issue Agent OS controller loop. Pulls issues from a GitHub issue queue (priority + dependency resolved), fans out triage to parallel triager agents, dispatches coder / reviewer / splitter / planner / debugger workers based on verdicts, drives review loops, merges approved PRs, and repeats until the queue is empty. Use when the user asks to "process issues", "run the issue controller", "work through the queue", or similar. Do NOT use for single-issue ad-hoc work — this is the full automated loop.

Manual-first workflow for designing, building, evaluating, and upgrading real agentic systems. Use when the job involves tool-using agents, agent orchestration, agent runtimes or platforms, multi-agent systems, framework selection, evals, or guardrails; do not use for prompt-only work or ordinary implementation planning.

Manual-first brainstorming workflow for turning ambiguous ideas or competing directions into an approved decision before planning or implementation. Activate when the user explicitly asks to brainstorm, explore options, compare approaches, or pressure-test a direction. Do not activate for clarification, review, detailed planning, or straightforward execution once a direction is already chosen.

Hybrid workflow for turning ambiguous implementation requests into actionable scope before execution. Use explicitly when the user asks to clarify requirements, disambiguate a request, or spec a change before building it. Activate automatically only when an implementation-oriented request contains direct contradictions, missing acceptance criteria that would materially change the build, or an explicit request to proceed with documented assumptions. Do not activate for open-ended direction-finding, brainstorming, or already-specific actionable tasks.

Manual-first architecture methodology for system design and technology selection using problem-first framing, stack-agnostic trade-offs, uncertainty handling, and future-aware decision lenses. Use when a project needs architecture direction or selection decisions; do not use for pure implementation planning, narrow stack research, or review-only tasks.

Use when the user explicitly asks about `git worktree`, wants a parallel checkout or isolated branch work without stashing or recloning, needs help resolving a branch that is already checked out elsewhere, or wants to inspect, compare, clean up, or repair worktrees safely. Do not use for ordinary single-branch Git help, repo bootstrap tasks, or broader advanced-Git sessions where worktrees are only incidental.

Manual-first heavy planning protocol for complex technical implementation work. Use when the user or the planner agent explicitly needs a deep executable plan for a feature, refactor, migration, or architecture change after the direction is already chosen. Do not use for brainstorming, choosing between competing directions, requirements clarification, business planning, generic project management, or code execution.

Govern a repository's `docs/` directory or docs site by reorganizing structure, consolidating overlapping pages, refreshing stale docs after code changes, and auditing documentation for drift, dead links, duplication, and missing source-of-truth boundaries. Use when the user explicitly wants to clean up, restructure, maintain, or verify a docs set rather than write a README or review a PR. Do not use it for README-only requests, one-off root-level docs outside the docs set, deep external research, or generic PR/diff review.

Structured review skill for PRs, diffs, commits, staged changes, code, documentation, API specs, database migrations, infrastructure config, and design documents. Activate for explicit, artifact-scoped review requests such as reviewing a PR, diff, commit, branch, staged changes, or named files, or phrases like "code review", "review my changes", "find bugs in", or "security review". Produces severity-graded findings with content-aware analysis, behavioral-change checks, and actionable fix directions. Do NOT activate for brainstorming, open-ended design feedback, governance/library audits, or general conversations about whether an approach seems good.

Manage the lifecycle of agent skills: discover patterns, create or update a skill, validate it, evaluate it, project or install it, or audit a skill library. Use only when the request is explicitly about skills or skill libraries, not for general code tasks or agent management.

Create, update, or review a repository README when the user explicitly asks for README work. Use it to draft a new README, refresh an existing README after project changes, or audit a README against the current repository. Do not use it for general documentation, API docs, architecture docs, or documentation tasks that do not specifically target a README file.

Use this skill for substantial business planning work: drafting or revising a business plan, investor-ready financial analysis, market sizing, pitch deck narrative, strategic review, or business valuation. It is especially useful when the user needs structured commercial thinking, investor-grade outputs, China fundraising context, or AI/agent business analysis. Prefer this skill when the main deliverable is a business analysis or plan, not open-ended idea exploration, general current-events research, or final file production in slides, docs, or spreadsheets.

Use when the user explicitly asks for comprehensive, citation-backed research such as a deep dive, due diligence, market analysis, or a multi-source comparison/report. Do not activate for quick factual lookups, ordinary coding tasks, or routine content generation unless the user first asks for research or source verification.

Manual-first workflow skill for redesigning, debugging, hardening, or productionizing LLM prompts and system prompts. Use when the user needs end-to-end prompt work such as stabilizing behavior, migrating across models, or preparing a prompt for production. Do not use for quick technique lookup or for general LLM-powered coding tasks.

Research existing community skills for a skill topic, pause for candidate confirmation, and produce a Fusion Report for downstream skill creation. Use only for skill-ecosystem research, not installation, general research, or direct skill authoring.

Comprehensive editorial review for Chinese content before publication. Use when the user asks to 'review my article', 'proofread this', 'check for errors', 'editorial review', 'pre-publish check', 'compliance review', or any Chinese text that needs proofreading, grammar checking, fact verification, political sensitivity screening, or advertising law compliance before release. Also activate when the user says '审校', '校对', '审核', '检查错误', '发布前检查', '合规审查', or submits Chinese content asking for quality review. Covers articles, news, blog posts, press releases, marketing copy, social media posts, and any public-facing Chinese text. Even if the user only asks for 'a quick check' or 'look this over', use this skill to ensure thorough coverage.