メニュー
jleechanbrain / Hermes-specific harness — gateway, canary, deploy, lane backlog, staging vs prod. Use with user-scope harness-engineering.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when dispatching work through the Hermes gateway with /claw, especially when the task may resolve slash commands or hand off into AO worker orchestration.
12-hour autonomous evolution loop — observe, measure, diagnose, fix, repeat every 10min. Drives zero-touch rate up by finding friction and dispatching fixes.
Custom evolve loop for jleechanbrain orchestrator — drains dropped Slack thread backlog via /claw, fixes hermes issues, proposes new work items. Max 50 items, newest-first.
Delegate coding change tasks to Agent-Orchestrator (AO). Triggered by the keyword "agento" anywhere in a message, or by default for any coding task. Never use mctrl unless explicitly requested.
Spawn an Agent-Orchestrator (AO) worker that uses the minimax CLI with the M3 model (via the minimax Anthropic-compatible API). Use when: the task should be dispatched to a worker that uses minimax M3 specifically, when the user says 'use minimax', 'use the M3 model', 'use the minimax worker', 'ao spawn minimax', or 'minimax CLI for AO'. This skill verifies the env (MINIMAX_MODEL, MINIMAX_API_KEY, ANTHROPIC_BASE_URL) end-to-end, then runs `ao spawn --agent minimax`. Verified 2026-06-13 against session ao-6355 / PR #678.
Enforce headless browser automation by default for Hermes — Playwright MCP and superpowers-chrome. Use when opening browsers, scraping, UI tests, localhost verification, or any chrome_use_browser / Playwright call. Never headed unless Jeffrey explicitly requests visible browser.
| name | hermes-harness |
| description | jleechanbrain / Hermes-specific harness — gateway, canary, deploy, lane backlog, staging vs prod. Use with user-scope harness-engineering. |
Scope: This skill applies to ~/.smartclaw (the jleechanbrain harness repo) and production/staging gateway operations. It does not replace the general protocol in ~/.claude/skills/harness-engineering/SKILL.md — it extends it for Hermes-specific failure modes.
curl /health is 200lane wait exceeded / queueAhead in gateway.err.log~/.smartclaw/ vs ~/.smartclaw_prod/)~/.claude/skills/harness-engineering/SKILL.md and ~/.claude/commands/harness.mdCLAUDE.md — Gateway (Local Machine), WS churn, session locks, lane backlogagent-orchestrator.yaml — reactions, evolve loop references| Symptom | Where to look | Harness fix layer |
|---|---|---|
Slack silent, /health OK | gateway.err.log: lane wait exceeded, queueAhead | CLAUDE.md, doctor.sh + staging-canary.sh lane probe |
session file locked | Stale .jsonl.lock in $HERMES_HOME/agents/main/sessions/ | Lock cleanup + restart (see CLAUDE.md) |
| Staging missing, prod OK | launchctl print gui/$UID/ai.smartclaw.staging | Bootstrap staging plist |
| Multi-gateway orphan | pgrep -x hermes > expected | deploy.sh / CLAUDE.md single-instance |
| WS churn / pong starvation | SlackWebSocket:N high | Lower timeoutSeconds × maxConcurrent per CLAUDE.md |
bash scripts/doctor.sh (or HERMES_DOCTOR_SKIP_INFERENCE=1 when appropriate)bash scripts/staging-canary.sh --port 8643 for prod config path when testing prod (includes gateway.err.log lane-backlog probe; parity with doctor)/health alone cannot prove Slack delivery; combine with steps 1–2./harness/harness (~/.claude/commands/harness.md) — 5 Whys, failure classes, audit mode — any repo./harness (.claude/commands/harness.md in this workspace) — short pointer + Hermes gates; wins for project-specific detail when both exist.curl /health without checking gateway.err.log tail~/.smartclaw/ directly in production without deploy flow (see CLAUDE.md worktree rule) — harness fixes still go through PR unless emergency exception