Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/teamspwk/nova/audit-selfgit clone --depth 1 https://github.com/TeamSPWK/novaWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00066 | $0.03363 |
| Opus 5 | $0.00033 | $0.01682 |
| Sonnet 5 | $0.00013 | $0.00673 |
| Haiku 4.5 | $0.00007 | $0.00336 |
Grade A, and why
audit-self scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 268 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/nova:audit-self
Nova 플러그인 자기 보안 진단 — 정적 분석 기반.
사용법
/nova:audit-self # 전체 5 카테고리 스캔 (~30K 토큰)
/nova:audit-self --category hooks # 단일 카테고리만 (~6K 토큰)
/nova:audit-self --jury # Red/Blue/Auditor 3에이전트 적대적 검증 (~50K 토큰, v5.23.0+)
비용
평균 1회 실행 ~30K 토큰 — 룰셋 30+ 룰 + 검사 대상 ~6900줄. --category {plugin|hooks|agents|skills|commands} 옵션으로 ~6K 토큰까지 축소 가능. 자주 호출 시 카테고리 분할 권장.
scan_targets
Phase 1 스캔 대상 — Glob 패턴으로 로드한 후 룰별 condition이 매칭되는 파일만 조건부 Read.
.claude-plugin/plugin.json— 플러그인 매니페스트hooks/*.sh— 훅 셸 스크립트agents/*.md— 에이전트 정의skills/*/SKILL.md— 스킬 정의commands/*.md— 슬래시커맨드 정의
exclusion_list
검사자/검사 대상 분리 원칙 — 메타-루프 가드 (R1 완화). 검사자(security-engineer)가 자기 정의를 검사하면 자가 합리화 편향이 발생하므로 명시 제외한다.
agents/security-engineer.md← 검사자 자기 정의commands/audit-self.md← audit-self 커맨드 자기 정의 (메타 루프 회피)docs/security-rules.md← 룰셋 자체 (R10 공급망 무결성은 별도 처리, v5.23.0)
메타 노트: 검사자/검사 대상 분리 원칙이 깨지면 결과 무효. v5.23.0의
--juryRed/Blue/Auditor 위임으로 자기 검사를 외부화 예정.
Phase 1: 카테고리 스캔 (룰셋 로드 + 파일 Glob)
docs/security-rules.md로드 → 룰 스키마 파싱 (7 필드 — id/category/severity/condition/normal_example/risk_example/mitigation)--category {X}지정 시 해당 카테고리 룰만 활성. 미지정 시 전체 활성- scan_targets Glob 결과 수집 → exclusion_list 항목 제거
- 각 룰의
condition1-liner를 활성 파일에 적용 (Read 도구로 조건부 로드)
Phase 2: security-engineer 호출
Agent 도구로 nova:security-engineer 서브에이전트를 spawn (Read/Glob/Grep만 — Edit/Write/Bash disallowed).
전달 컨텍스트:
mode: self-audit
scan_targets: <Phase 1 결과>
exclusion_list: <exclusion_list 항목>
active_rules: <카테고리 필터 적용된 룰 목록>
rules_doc: docs/security-rules.md
서브에이전트는 agents/security-engineer.md 의 "Nova 자기 코드 감사 모드" 섹션 규약을 따라 마크다운 리포트를 반환한다 (Critical/Warning/Info 분류).
Phase 2.5: --jury 분기 (v5.23.0+, ECC §P2-3 흡수)
--jury 플래그가 지정된 경우, Phase 2의 단일 security-engineer 결과를 3 페르소나 적대적 검증으로 보강한다. skills/jury/SKILL.md 모드 3 (보안 진단) 호출.
전달 컨텍스트:
mode: audit
target: nova-self-codebase
rules_doc: docs/security-rules.md
scan_targets: <Phase 1 결과>
exclusion_list: <exclusion_list 항목>
security_engineer_report: <Phase 2 결과>
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 268 lines · 66 tokens per session scan A dda4892a01c3
audit-self is a command published in the GitHub repository TeamSPWK/nova (2 stars, last pushed 2d ago), licensed MIT. It adds 66 tokens to every session and 3,363 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
loopx
Run the loopx setup helper and read its output.
grow-scanner
Autonomous task: increase real-world adoption of the AgentDiscover scanner.
gedd-chat
You are a GEDD coaching assistant. You guide the user through building a golden evaluation dataset for their AI agent using Open Coding methodology, then help them evaluate and annotate responses — all without leaving Claude Code.
gedd
You are a GEDD coaching agent. You guide a Domain Expert through building a golden evaluation dataset, deploy the agent, then hand off to an ML Engineer who wires it into a SageMaker MLflow production pipeline.
gedd-status
Read session.json (use the Read tool) and display a concise dashboard. If the file doesn't exist, say so and suggest running /gedd to start.
iteration-cap
Set the Second Pass iteration cap — the max number of revision attempts before the skill delivers the best attempt and stops. Accepts a positive integer.