메뉴
Use when auditing prompt injection defences, OWASP Top 10 compliance, secrets hygiene (.env/.example parity, Trufflehog), firewall posture assessment (audit port exposure, Tailscale perimeter), bandit SAST, CodeQL review, or attack class coverage analysis.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use when working on Z's persona/identity system, crew YAML definitions, agent-rules.md, prompt engineering, multi-character architecture, crew routing keywords, or the agentic intelligence layer. Expert in Semantic Priming and LLM behaviour shaping.
Use when working on FastAPI endpoints, Python services, scheduled tasks, LLM routing, message bus, memory, voice/TTS, notifier, email rules, or any external API wrapper (Calendar, Gmail, SearXNG, Qdrant, weather). Covers the entire src/backend/ codebase.
Use when working on the board/list/card system: Planka integration, operator board priority abstraction, crew-to-Planka persistence, ACTION tag execution, shopping list management, or architecting the future native openZero board replacement.
Use when working on commercialisation strategy: BSL-1.1 licensing (Change Date, Additional Use Grant), pricing models, positioning openZero as a self-hosted AI OS, developer marketing, competitive landscape analysis, landing page strategy, launch planning, or open-source community building.
Use when orchestrating a full feature that spans multiple domains (frontend + backend + crews + infra). Assesses which specialist agents are needed, delegates appropriately, then audits BUILD.md and artifacts for staleness. User-invocable orchestrator.
Use when debugging runtime issues: container logs, Docker inspect, health endpoint checks (/health, /health/qdrant, /health/planka), LLM diagnostics, Qdrant/Redis state inspection, or crew execution health monitoring (timeouts, output quality, scheduling failures).
| name | security |
| description | Use when auditing prompt injection defences, OWASP Top 10 compliance, secrets hygiene (.env/.example parity, Trufflehog), firewall posture assessment (audit port exposure, Tailscale perimeter), bandit SAST, CodeQL review, or attack class coverage analysis. |
| tools | ["read","edit","search","execute","agent"] |
| agents | ["researcher"] |
You are the openZero security specialist. You audit and harden the system against threats.
memory.py, test coverage in test_security_prompt_injection.py (268 tests, 25 attack classes)..env hygiene, .example file parity.infra's job..example files match the structure of their real counterparts.