Set up or diagnose Harness project configuration, including project initialization, CI, Codex CLI, memory, agents, and skill mirrors. Choose a focused setup area, follow its reference guidance, and check the resulting configuration where needed.
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.
Run score Not tested
See how to installReview the instructions the agent will receive before you install.
Harness の統合セットアップスキル。 以下の旧スキルを統合:
setup — 統合セットアップハブharness-init — プロジェクト初期化harness-update — Harness アップデートmaintenance — ファイル整理・クリーンアップ| サブコマンド | 動作 | 詳細 |
|---|---|---|
/harness-setup init | 新規プロジェクト初期化（CLAUDE.md + Plans.md + hooks + sync + doctor） | ${CLAUDE_SKILL_DIR}/references/init.md |
/harness-setup ci | CI/CD パイプライン設定 | ${CLAUDE_SKILL_DIR}/references/ci.md |
/harness-setup codex | Codex CLI インストール・設定 | ${CLAUDE_SKILL_DIR}/references/codex.md |
/harness-setup harness-mem | harness-mem 統合・メモリ設定 | ${CLAUDE_SKILL_DIR}/references/harness-mem.md |
/harness-setup mirrors | skills/ → 公開 mirror bundle 更新 | ${CLAUDE_SKILL_DIR}/references/mirrors-agents-localize.md |
/harness-setup agents | agents/ エージェント設定 | ${CLAUDE_SKILL_DIR}/references/mirrors-agents-localize.md |
/harness-setup localize | CLAUDE.md ルールのローカライズ | ${CLAUDE_SKILL_DIR}/references/mirrors-agents-localize.md |
| marketplace / update | Plugin install, update, dependency policy | ${CLAUDE_SKILL_DIR}/references/marketplace.md |
| maintenance | ファイル整理・クリーンアップ | ${CLAUDE_SKILL_DIR}/references/maintenance.md |
Built-in slash discovery (CC 2.1.108+):
/initのような built-in slash command も発見される。 Harness 固有の bootstrap が必要な時だけ/harness-setup initと使い分ける。
Claude Code setup guidance (CC 2.1.120+): MCP
alwaysLoad、${CLAUDE_EFFORT}、claude plugin prune、claude project purge、ANTHROPIC_BEDROCK_SERVICE_TIER、claude_code.skill_activated.invocation_trigger、 Windows PowerShell primary shell、forked skills / subagents の deferred tools はdocs/claude-code-setup-mcp-telemetry-provider.mdを正本として扱う。
Codex plugin workflows: Codex
/goalとPlans.mdを二重管理しない。 plugin-bundled hooks は opt-in、external agent import は ownership 明記、 MultiAgentV2 /agents.max_threads = 8は上限として扱い、 sticky environments / app-server artifacts は safe default を優先する。 Codex0.130.0stable のcodex remote-control、large thread pagination、 selected-environmentview_image、live app-server config refresh、 accurate turn diffs、plugin details bundled hooks、sharing discoverability controls はdocs/codex-plugin-workflows-policy.mdを正本として扱う。 詳細はdocs/codex-plugin-workflows-policy.mdと${CLAUDE_SKILL_DIR}/references/codex.mdを参照。
状態確認だけの依頼は read-only の診断で完了する。setup 依頼は対象、望む状態、合格条件を既存設定と照合し、明示された surface の変更を進める。 不足情報は設定や契約を読み取りで回収する。任意機能を追加したり、対象外の設定や保護された設定を目的から推測して変更したりしない。
${CLAUDE_SKILL_DIR}/references/*.md を読む。harness doctor、bash scripts/sync-skill-mirrors.sh --check、bash scripts/ci/check-consistency.sh で確認する。docs/plugin-managed-settings-policy.md — plugin managed settings policy; normal defaults must not inherit managed marketplace restrictions.docs/codex-provider-setup-policy.md — Codex provider setup policy, including amazon-bedrock and model_provider = "amazon-bedrock" examples.0.123.0 以降の MCP diagnostics / plugin MCP loading guidance: use /mcp verbose and docs/codex-mcp-diagnostics.md for diagnostics.docs/codex-sandbox-execution-policy.md for remote_sandbox_config and execution constraints.詳細手順は ${CLAUDE_SKILL_DIR}/references/cursor.md を読む。
契約アンカー: set-impl-backend.sh は AI 実行可。permissions.json、.cursorignore、*.cursor.sh allowlist はユーザー手動で、AI が編集できない protected path として扱う。
根拠ルールは .claude/rules/cursor-cli-only.md と docs/sandbox-allowlist-recipe.md。
${CLAUDE_SKILL_DIR}/references/init.md — init, Go binary verification, plugin sync, doctor.${CLAUDE_SKILL_DIR}/references/ci.md — GitHub Actions CI setup.${CLAUDE_SKILL_DIR}/references/codex.md — Codex CLI, provider/model metadata, app-server, MCP diagnostics, sandbox policy.${CLAUDE_SKILL_DIR}/references/cursor.md — Cursor implementation backend setup and boundaries.${CLAUDE_SKILL_DIR}/references/harness-mem.md — memory directory and template setup.${CLAUDE_SKILL_DIR}/references/mirrors-agents-localize.md — mirror sync, agent setup, localization.${CLAUDE_SKILL_DIR}/references/marketplace.md — plugin marketplace install/update and managed dependency policy.${CLAUDE_SKILL_DIR}/references/maintenance.md — cleanup commands and related skills.harness-sync — 設定/Plans/git 状態の同期確認harness-work — 実装タスク実行harness-review — 品質レビューmaintenance — ファイル整理（統合済み）No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Use the agent-browser CLI to navigate web pages, inspect interactive elements, fill forms, capture screenshots, and collect page information. It helps check web app behavior through browser actions and snapshots.
Run planned software tasks as a coordinated team workflow. It assigns ready tasks to isolated workers, performs per-task and integrated reviews, incorporates approved changes, and reports progress and completion. Worker execution can use the selected backend, including optional Cursor delegation.
Run planned work through a team workflow: delegate implementation to workers, coordinate independent review and revisions, then report the outcome. Select a task range or backend, or run all unfinished tasks in one command.
Review changes that integrate Claude Code or Codex updates. Classify each new Feature Table entry as implemented, inherited upstream, or tracked in Plans, and flag entries that are only documented without supporting work.
Diagnose failed CI runs by reviewing logs, reproduction details, specifications, and recent commits. Identify whether the cause lies in implementation, tests, or the environment, then make authorized fixes without hiding failures by skipping checks.
Send a question, investigation, or design discussion to cursor-agent in read-only ask mode. It returns a second opinion or evidence-based findings that the host agent summarizes, without making file changes or running implementation work.