메뉴
One-page operator cheatsheet for the Aufgaard plugin. The "what to check, when, why" reference. Auto-loaded so day-to-day routing is fast. Use as the lightweight summary when the user wants a quick recap of capabilities.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
| name | operator-cheatsheet |
| description | One-page operator cheatsheet for the Aufgaard plugin. The "what to check, when, why" reference. Auto-loaded so day-to-day routing is fast. Use as the lightweight summary when the user wants a quick recap of capabilities. |
| when_to_use | When the user asks "what can you do", "show me the cheatsheet", "what does Aufgaard do", or wants a one-page overview. Also use as your own internal reference when planning multi-step responses. |
| user-invocable | false |
| Surface | When | What you see |
|---|---|---|
| SessionStart hook | Every Claude Code session open | 🟢 Daily deployment overview — gateway uptime / plugins loaded / errors / yesterday's spend |
| PreToolUse(Bash) hook | Every shell command | bash-vet auto-vets; BLOCK-tier rejects before execution |
| PreToolUse(Edit on .claude/) hook | Writes to agent-config files | skill-vetter auto-vets; BLOCK-tier prevents commit |
| PostToolUse(Task) hook | Subagent task completion | output-vetter verifies the claim against actual state |
| Background monitors | Continuous | Push notifications inline: 429-imminent / silent-failure detected / gateway-down / regression-catalogue match / dangerous-skill landed |
/aufgaard:vet-bash <command> — vet a specific shell command/aufgaard:vet-skill <path> — vet a third-party skill before installing/aufgaard:vet-config <path> — vet agent-config files (CLAUDE.md, AGENTS.md, .cursor/rules)/aufgaard:cost-overview [period] — cost dashboard + 429 prediction/aufgaard:health-check [scope] — deployment vitals/aufgaard:should-i-upgrade <package> — upgrade safety advisor/aufgaard:silent-failures [period] — recent silent-failure detections/aufgaard:verify-claim "<claim>" — verify agent's stated outcome/aufgaard:find-swallowed-exceptions <path> — scan Python for silent error masking/aufgaard:production-audit — comprehensive 35-pattern audit; calls all 7 MCPs in parallel; one-page reportWhat's the symptom?
├── Cost / billing concern → /aufgaard:cost-overview
│ └── 429 imminent specifically → cost-overview shows 429 status inline
├── Cron / scheduled job concern → /aufgaard:silent-failures
├── Should I install this skill → /aufgaard:vet-skill
├── Did my agent's claim match reality → /aufgaard:verify-claim
├── Should I run this bash command → /aufgaard:vet-bash
├── Should I upgrade this package → /aufgaard:should-i-upgrade
├── Is my agent code error-handling solid → /aufgaard:find-swallowed-exceptions
├── Is my deployment healthy → /aufgaard:health-check
├── Just got a PR with .claude/ changes → /aufgaard:vet-config
└── End-to-end review → /aufgaard:production-audit
| Skill family | MCP | PyPI install |
|---|---|---|
| vet-bash | bash-vet-mcp | pip install bash-vet-mcp |
| vet-skill / vet-config | openclaw-skill-vetter-mcp | pip install openclaw-skill-vetter-mcp |
| cost-overview | openclaw-cost-tracker-mcp | pip install openclaw-cost-tracker-mcp |
| health-check | openclaw-health-mcp | pip install openclaw-health-mcp |
| should-i-upgrade | openclaw-upgrade-orchestrator-mcp | pip install openclaw-upgrade-orchestrator-mcp |
| silent-failures | silentwatch-mcp | pip install silentwatch-mcp |
| verify-claim / find-swallowed-exceptions | openclaw-output-vetter-mcp | pip install openclaw-output-vetter-mcp |
23 of 35 documented production-AI failure patterns are covered (see production-patterns skill for the full catalogue). 12 are uncovered — flagged honestly in audit reports. The uncovered patterns cluster in observability-completeness, deploy-config security, runtime-monitoring extensions, and cross-model deception. These are the next-build frontier.
The [Production-AI MCP Suite Bundle](https://temurah.gumroad.com/l/production-ai-mcp-suite) ($29) includes the 8-page Field Reference PDF — full 35-pattern catalogue with verbatim operator-incident sources, MCP coverage map, audit rubric, and 7-MCP cheat sheet.
SOC 직업 분류 기준
Show current production AI cost overview — totals, top spenders by agent, per-provider breakdown, anomaly detection, and time-to-429 rate-limit prediction. Use when the user asks "what's my AI bill", "which agent is burning tokens", "am I about to hit the rate limit", or wants a cost dashboard. Cross-provider — works for Anthropic, OpenAI, Gemini, Bedrock, Ollama.
Scan Python source files for swallowed-exception patterns that silently turn errors into fake successes. Catches bare `except` blocks that pass / return None / return mock objects, log-and-fake-success handlers, and mock-substitution-on-error. AST-based — not just regex. Use before any deploy of new agent code, on the working directory after a bug fix, or routinely on production-path Python files.
Show current AI deployment health overview — gateway status, plugin/skill registry, recent errors, CPU/RAM pressure, OOM history, cron status, disk pressure, upgrade outcome. The "vital signs" panel for a production AI deployment. Use when the user asks "is my deployment healthy", "what's wrong with my agent", "are services up", or wants an at-a-glance status check.
Comprehensive production-AI deployment audit against the 35-pattern catalogue. Calls all 7 MCP servers in parallel — bash-vet, skill-vetter, cost-tracker, silentwatch, health-mcp, upgrade-orch, output-vetter — and synthesizes a one-page report with critical findings, audit score, and remediation recommendations. Use when the user asks "is my AI deployment healthy?", "audit my production AI setup", "run a production audit", or wants an end-to-end review of their AI deployment.
Check whether upgrading a package / runtime / model version is safe. Looks up the user-driven regression catalogue (8+ entries from real field reports) AND runs provider-side regression detection (catches Anthropic-April-23-style silent reasoning-effort downgrades). Returns a recommended upgrade path with mitigations. Use before any significant package upgrade, model-version bump, or runtime change.
Show recent silent-failure detections from cron / scheduled jobs. Catches the textbook patterns — exit-0 with empty stdout, length anomalies (output dramatically shorter than baseline), retry storms, action-budget leaks. Use when the user asks "is anything silently broken", "did Friday's cron actually run", "are my scheduled jobs working", or after a downstream consumer reports stale data.
