claude-code-harness

executor が返した advisor-request.v1 に対して方針だけ返す非実行 advisor

sprint-contract と review artifact を基準に verdict を返す read-only reviewer

analyze、scaffold、update-state の 3 モードで足場構築を行う統合 scaffolder

Harness の標準チーム構成は 5 ロール。

実装、preflight 自己点検、検証、commit 準備を 1 タスク単位で進める統合ワーカー

Control browser like hands and feet. Navigate, fill forms, screenshot, bring it on. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include 'go to [url]', 'click on', 'fill out the form', 'take a screenshot', 'scrape', 'automate', 'test the website', 'log into', or any browser interaction request. Do NOT load for: sharing URLs, embedding links, screenshot image files.

Implements authentication and payment features using Clerk, Supabase Auth, or Stripe. Use when user mentions login, authentication, payments, subscriptions, or Stripe. Do NOT load for: general UI work, database design, or non-auth features.

Team execution mode — backward-compatible alias for harness-work with team orchestration.

Validates brainstormed ideas with Cursor PM, updates Plans.md, then handoff back. Cursor ↔ Claude Code 2-Agent workflow support. Use when user mentions Cursor PM handoff, 2-agent plan validation, CC-Cursor round trip, or brainstorm review. Do NOT load for: implementation work, single-agent tasks, or direct coding.

Quality guardrail for Claude/Codex update integration. Detects doc-only Feature Table additions and requires implementation or explicit planning. Internal use only.

CI red? Call us. Pipeline fire brigade deploys. Use when user mentions CI failures, build errors, test failures, or pipeline issues. Do NOT load for: local builds, standard implementation work, reviews, or setup.

Local-only skill for researching Claude Code and Codex upstream releases, selecting high-value adaptations, and implementing meaningful Harness upgrades. Not for distribution.

Auto-generate CRUD quickly. Boilerplate left to AI. Use when user mentions CRUD, entity generation, or wants to create API endpoints. Do NOT load for: UI component creation, form design, database schema discussions.

Controls session resume/fork(branch) for /work based on --resume/--fork flags. Updates session.json and session.events.jsonl. Internal workflow use only. Do NOT load for: user session management, login state, app state handling.

Initializes session with environment checks and task status overview. Use when user mentions starting a session, beginning work, or status checks. Do NOT load for: implementation work, reviews, or mid-session tasks.

Manages cross-session learning and memory persistence. Use when user asks about previous sessions, history, or to continue from before. Do NOT load for: implementation work, reviews, or ad-hoc information.

HAR: Sync Plans.md with implementation. Drift detect, marker update, retrospective. Trigger: sync-status, where am I, check progress. --snapshot for snapshots. Do NOT load for: planning, implementation, review, release.

HAR: Execute Plans.md tasks from single task to full parallel team run. Trigger: implement, execute, do everything, breezing, team run, parallel. Do NOT load for: planning, review, release, setup.

File cleanup and archiving. Tidies up bloated Plans.md, session-log.md, old logs, and state files. Trigger: /maintenance, cleanup, archive, organize, split session-log. Do NOT load for: implementation, review, release, new feature development.

Manage SSOT, memory, and cross-tool memory search. Guardian of decisions.md and patterns.md. Use when user mentions memory, SSOT, decisions.md, patterns.md, merging, migration, SSOT promotion, sync memory, save learnings, memory search, harness-mem, past decisions, or record this. Do NOT load for: implementation work, reviews, ad-hoc notes, or in-session logging.

Generate NotebookLM YAML and slides. Document craftsman shows skill. Use when user mentions NotebookLM, YAML, slides, or presentations. Do NOT load for: implementation work, code fixes, reviews, or deployments.

Provides development principles, guidelines, and VibeCoder guidance. Use when user mentions principles, guidelines, safety, or diff-aware editing. Do not use for actual implementation—use the impl skill instead.

HAR: Multi-angle code, plan, scope review. Security/quality check. Trigger: review, code review, plan review, scope analysis. Do NOT load for: implementation, new features, bugfix, setup, release.

Deploy to Vercel/Netlify. One-way ticket to production arranged. Use when user mentions deployment, Vercel, Netlify, analytics, or health checks. Do NOT load for: implementation work, local development, reviews, or setup.

Generate project intro slides with Nano Banana Pro. Use when user mentions slide, project slide, 1-page summary, or visual introduction.

Auto-generate product demo videos. A picture worth thousand words, embodied. Use when user mentions '/generate-video', video generation, product demos, or visual documentation. Do NOT load for: embedding video players, live demos, video playback features. Requires Remotion setup.

Use gogcli for Google Workspace CLI operations (Drive/Sheets/Docs/Slides). Trigger when a user asks to check, list, search, export, read, or update Google files via gogcli; when a Google URL/ID needs parsing; when auth/account selection or safe read-only workflows are needed; or when troubleshooting gogcli access/errors. Do NOT load for: general file operations, non-Google cloud storage, or standard shell commands.

Long-running task loop using /loop (Claude Code dynamic mode) and ScheduleWakeup to re-enter with fresh context on each wake-up. Internally invokes harness-work through Agent. Trigger: long-running, loop, wake-up, autonomous. Do NOT load for: one-shot task execution, review, release, planning.

HAR: Task planning, Plans.md management, progress sync. Trigger: create a plan, add tasks, update Plans.md, mark complete, check progress. Do NOT load for: implementation, review, release.

Generic release automation for projects using Keep a Changelog + GitHub. Single confirmation gate then end-to-end automation: bump detection, CHANGELOG promotion, tag, GitHub Release. Trigger: release, version bump, publish. Do NOT load for: implementation, review, planning, setup.

HAR: Project init, tool setup, agent config, memory setup, skill mirror sync. Trigger: setup, init, new project, CI/Codex setup, harness-mem, mirror. Do NOT load for: implementation, review, release, planning.

Manages session state transitions per SESSION_ORCHESTRATION.md. Controls state updates at /work phase boundaries, escalated transitions on error, and initialized restoration on session resume. Internal workflow use only. Do NOT load for: user session management, login state, app state handling.

Unified session management window. Handles initialization, memory, state all-in-one. Explicit /session invocation only — sub-skills handle auto-delegation. Use when managing Claude Code sessions, /session command. Do NOT load for: app user sessions, login state, authentication features.

Generates UI components and feedback forms. Use when user mentions components, UI, hero sections, forms, feedback, or contact requests. Do NOT load for: authentication features, backend implementation, database operations, or business logic.

Guides VibeCoder (non-technical users) through natural language development. Use when user asks what to do next, how to use the system, needs help, or is stuck. Do NOT load for: technical-user work, direct implementation requests, or reviews.

Provides guidance on Cursor ↔ Claude Code 2-agent workflow. Use when user asks about workflow, collaboration, or process. Do NOT load for: implementation work, workflow setup, or executing handoffs.

59 hooks across 26 events

Plan/Work/Review ワークフローに最適化された構造化出力スタイル。進捗追跡とフェーズ別の出力形式を提供。

Auto-arm session monitoring for harness-mem health, advisor/reviewer state, and Plans.md drift.