apm-workspace

Route work in the `~/.apm` global APM workspace: decide what owns a change, which path is the source of truth, and which APM rollout / `mise` task to run. Use for skill update and redistribution requests (`skillアップデートして再配布して`, `再配布`), `apm.yml` / `apm.lock.yaml` and managed catalog rollout, manual-skills package state, orphaned APM packages, checked-out external dependency repositories, optional repository-scoped skill packages, and `apmのバージョンあげて` / pinned `apm` source checks. For skill body, description, script, reference, or asset design itself, coordinate with `skill-creator`; for general mise usage outside the APM workspace (including `mise upgrade TOOL` and `minimum_release_age`), use `mise`.

Use when deciding whether to install, enable, or troubleshoot Headroom context compression for local AI agent workflows, especially MCP, proxy, or agent-wrap usage.

Use when adding or reviewing CI-stability hooks and gates: Lefthook/pre-commit/pre-push configuration, aligning local push gates with GitHub Actions, deciding commit-vs-push gate scope, or migrating pre-commit to Lefthook. For mise task design itself, use `mise`; for debugging failing PR checks, use `gh-fix-ci`.

Curate external agent collections into this APM workspace. Use when evaluating whether an outside agent repository can be installed through APM, importing selected agents into catalog/agents, recording provenance, or refreshing curated agents from upstream.

Use when creating, rebuilding, or updating a Google Forms survey from a Markdown survey spec such as SURVEY_QUESTIONS.md, especially when the form should stay synchronized with question text, required flags, Google Forms item types, scale labels, and removed-question notes.

Slack app management knowledge for app settings, manifests, OAuth installs, scopes, bot users, Marketplace settings, App Home, display names, sender names, mrkdwn message formatting, and safe workspace-visible changes. Use when diagnosing or changing a Slack app, installed app, bot user, app label, DM header, chat sender, message text decoration, Slack links, block quotes, app approval/install state, token impact, Slack API app settings, Slack Marketplace settings, or stale Slack app name propagation.

Use when planning, requirements, design, estimation, or task decomposition needs pre-implementation failure prediction, blind-spot discovery, planning review, design validation, or answers to "what could go wrong" / "what am I missing". Use before implementation begins.

[What] Guide for creating Claude Code rules. Provides guidance on when to use CLAUDE.md vs .claude/rules/ and step-by-step creation procedures. [When] Use when: you want to enforce project standards, guide AI behavior, or create/modify rules. [Keywords] rules, guidelines, project standards, policy, enforcement, claude.md

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

Analyzes Claude Code session transcripts to evaluate skill portfolio health — routing errors, attention competition between descriptions, and coverage gaps. Generates an interactive HTML report with per-skill health cards, competition matrix, attention budget analysis, and actionable patches. Unlike skill-creator which optimizes individual skills in isolation, skill-auditor optimizes the portfolio as a system, detecting cross-skill attention theft and cascade risks. Use when user says "audit my skills", "skill audit", "run skill-auditor", "analyze skill routing", "check skill competition", "portfolio health", "スキル監査", "スキルの精度を分析", "スキルルーティング分析".

Use when creating or editing herdr configuration (`~/.config/herdr/config.toml`) — keybindings, themes, UI/sidebar, notifications, sounds, terminal defaults, worktrees, or experimental options — when herdr shows a startup config warning, or when inspecting or controlling live Herdr workspaces, tabs, and panes through the CLI, including launching an interactive command or TUI in a specific tab. Do not use for generic terminal multiplexer comparisons.

Guidance for mise (mise-en-place) as task runner, tool version manager, and package manager. Use when working with mise.toml, task definitions and dependency graphs (`mise run`, depends/run), `[task_config].includes` and DB/env/dotenvx/secrets/tools task-family splitting, tool/package centralization (npm:/pipx:), `mise upgrade` and `minimum_release_age`, Windows settings (`run_windows`, `config.windows.toml`), `mise skills add`, `mise bootstrap` and its config sections (`[bootstrap.packages]`, `[dotfiles]`, launchd/systemd units), `mise dotfiles apply`, or migrating a Brewfile to mise bootstrap. For `~/.apm` rollout work (apm.yml, lockfile, which APM task to run), coordinate with `apm-usage`. For Home Manager / Nix Flake dotfiles, use `nix-dotfiles`.

[What] Specialized skill for systematic code quality improvement. Provides Phase 1→2→3 workflow for ESLint error fixing, type safety improvements, and code quality enhancements [When] Use when: users mention "ESLintエラー", "大量修正", "段階的修正", "code quality", or need systematic code quality improvement [Keywords] ESLintエラー, 大量修正, 段階的修正, code quality

変更ファイルを論理的な最小単位でグループ化し、git log のスタイルに合わせた Conventional Commits 形式でグループごとに個別コミットする。ユーザーが commit / push の実行（「コミットして push」`commit and push` など）を依頼したとき、「最小単位でコミット」`atomic commit` を求めたとき、または dotenvx-managed `.env.*` を含むコミット計画で使用する。push 単独、PR 作成、ブランチ作成、GitHub 公開は扱わず、コミット分割とメッセージ作成に範囲を絞る。

Run configurable local code reviews with project-type detection and tech-stack-specific criteria. Use when the user asks to review code or a diff (staged, recent commit, or branch), requests a quality assessment or star-rated evaluation, or wants a review that integrates project-specific guidelines. Targets local changes, optionally addresses open PR comments, and reports results in Japanese.

Review whether project entrypoint documentation needs updates before editing it. Use when the user asks to check docs drift or update docs after implementation, refactoring, configuration, operational workflow, UI rule, or design-system changes, or mentions CLAUDE.md, AGENTS.md, llms.txt, TODO.md, DESIGN.md, DESIGN_REVIEW.md, README.md, CHANGELOG.md, or docs/**. Default to proposing update candidates first instead of editing immediately.

Use when reviewing, validating, creating, updating, or fixing project documentation, especially docs directories and Markdown files governed by metadata, OKF / YAML frontmatter, tag, link, and size rules such as `.docs-manager-config.json`. Also use when a user asks to make docs OKF-compatible or says to use OKF docs, even if they do not spell out the frontmatter fields.

Use when managing Git worktrees with `git wt` or `git worktree`, especially for branch isolation, `.worktrees/` layout, parallel work, or diagnosing worktree create/remove errors.

Run Jina Reader-first public web research with cited synthesis. Use when the user asks to research current public web signals, search X/Twitter via Jina, compare public sources, or produce a concise evidence-backed research brief using s.jina.ai.

Use when the user wants help deciding, writing, or automatically starting a `/goal`; gives rough intent, implementation/fix/proceed instructions, incremental next-step instructions, issue text, plans, TODOs, review notes, or asks what the practical landing point should be.

Run a managed fix-and-review loop for complex implementation work: parallel subagent investigation feeding a main-session task backlog, evidence-backed triage, and repeated scored reviews until ship quality. Use when the user asks for a review fix loop / fix-and-review loop, asks to coordinate subagents with a task backlog and repeated quality review, or invokes this skill with a compact prompt whose conditions must be recovered from prior conversation and nearby repository context.

Use when operating dotenvx-managed environment files, encrypted `encrypted:` values, `.env.*` loading, `dotenvx run`, or when dotenvx/mise environment injection may affect CLI tools such as AWS CLI, Terraform, CDK, GitHub CLI, or perman-aws-vault. Trigger for `.env.development`, `.env.staging`, `.env.production`, `.env.dev`, env drift, secret redaction, and commands that need a clean environment instead of decrypted app env.

[What] Integrated refactoring workflow for TypeScript/JavaScript/React: similarity-ts (duplicate detection), react-doctor (React diagnostics), tsr (dead code removal), and boundary ownership scanning, orchestrated with code-quality-improvement. [When] Use when users mention リファクタ / refactor, 重複コード / duplicate code, コード整理 / cleanup, デッドコード・未使用ファイル・ 未使用 export removal, 共通 helper extraction, validation / Result / repository boundary, folder ownership, or cleanup-only plans. Do not use for feature implementation or task execution; use the normal implementation flow or executing-plans instead.

Runtime assets (codex-companion.mjs, prompts, schema) backing the catalog's /codex:* commands. Not directly invocable.

日本語の技術文書・書籍原稿の文章規範。整形（一文一行、引用ブロック、脚注、コラム記法）、段落と論証の構成（パラグラフライティング）、論証の厳密さ（ツッコミどころの除去）、読み手の負荷の管理、視点と語り、演出の抑制、LLM っぽい空句の禁止、冗長の排除を定める。日本語で技術書の章、草稿、記事、解説文を書くとき、または推敲・リライトするときに使用する。

Create, update, and review a paired `DESIGN.md` and `DESIGN_REVIEW.md` workflow for AI-assisted UI work. Use when Codex needs to extract a durable design system from existing code, screenshots, URLs, or rough product intent; decide whether guidance belongs in `DESIGN.md`, shared UI, or feature-local code; enforce strict Stitch-style `DESIGN.md` boundaries; or route review logic, exceptions, and escalation rules into `DESIGN_REVIEW.md`.

Use when scanning a repository to create, update, or clean up a repo-local `apm.yml` and install appropriate local APM skills or MCPs. Use for requests like `repo に apm.yml を置いて`, `この repo におすすめ skill 入れて`, `ローカルスキル整理`, `repo-local APM bootstrap`, or moving infra/runtime-specific skills out of global APM. Coordinate with `apm-usage` for global workspace ownership, lockfile rollout, and deployed target rules.

Use when deciding whether to install, keep, move, or remove MCP servers, or when setting up, reviewing, or troubleshooting MCP servers, especially for global vs repo-local placement, trend checks, config file locations, transport choices, credential handling, or startup failures.

Select and run expert review lanes for UI, forms, design specifications, accessibility, and multi-device product quality when ordinary /review or code review is too shallow.

Create or update audit-friendly architecture boundary documentation for a codebase. Use when documenting architecture boundaries, layer maps, folder or type responsibilities, Result/ServerAction boundaries, data flow, error flow, side effects, dependency direction, verification maps, forbidden crossings, or agent-facing architecture maps from repository evidence.

Maintain work-log artifacts and produce manager-facing Slack reports with business-day counts. Use when asked to update work logs, summarize work periods, split work-log files, compute business days, or prepare short manager reports from evidence such as git log, changelog, TODOs, issues, Slack notes, or existing work-log files.

Use when auditing noisy test, build, lint, package-manager, or CI commands before reducing output; deciding between tool-native quiet/silent/minimal options and optional agent-side compression; proposing changes that keep failures visible while suppressing successful boilerplate. Trigger when the user asks to audit noisy commands, quiet logs, reduce context, silence successful tests, or make CI less noisy. Do not edit command definitions unless the user explicitly approves the proposed change.

Review a pasted or freshly generated implementation, operation, research, or rollout plan as a pre-execution gate without executing it; return a concise human summary, prioritized plan-review findings, and a copy-ready prompt for another agent, with conditional routing for UI/frontend, superpowers, and subagent-suitable work.

Review UI changes for design-system compliance, reusable component routing, token usage, and interaction consistency.

Route and synthesize cross-source research across web, GitHub, social platforms, video, and community sources. Use when a research task may require multiple source types, source-routing decisions, Agent Reach, Jina Reader, or delegated web research.

Use when working with 1Password CLI (`op`) to find, inspect, create, edit, or update 1Password items such as logins, secure notes, API credentials, service account token items, repo-specific dotenvx `.env.keys` file attachments, and related secrets. Default to the Personal vault unless the user names another vault, use dotenvx-managed `OP_SERVICE_ACCOUNT_TOKEN` when provided for authentication, search before changing items, avoid exposing secret values, and require confirmation for ambiguous or destructive operations.

Use when a TypeScript or JavaScript codebase needs duplicate-code detection, similar function or type-definition analysis, or refactoring-candidate triage based on `similarity-ts`. Do not use for Python, Go, Rust, or general multi-language clone detection.

Use when a TypeScript or React codebase needs dead-code detection or safe unused-code removal with TSR, especially for unused exports, unused files, `.tsrignore`, and gradual cleanup workflows.

Code quality assurance with automatic fix iteration. Execute lint/format/test and automatically fix errors until all pass (max 3 attempts). [What] Detect project configuration (mise.toml, package.json), run format/lint/test, and iteratively fix errors [When] Use when: users mention "polish", "品質保証", "lint fix", "format fix", "quality check", or need automated code quality improvement with repeated fixes [Keywords] polish, quality assurance, lint, format, test, automatic fix, code quality

Predictive code analysis for identifying potential risks, anti-patterns, and future maintenance issues. Use when assessing code health or planning refactoring.