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 skills/cafitac/hermit-agent/hermit_auditnpx skills add cafitac/hermit-agent --skill hermit_auditgit clone --depth 1 https://github.com/cafitac/hermit-agentWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/cafitac/hermit-agent/hermit_audit)<a href="https://agentmods.dev/skills/cafitac/hermit-agent/hermit_audit"><img src="https://agentmods.dev/badge/skills/cafitac/hermit-agent/hermit_audit.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00133 | $0.01640 |
| Opus 5 | $0.00067 | $0.00820 |
| Sonnet 5 | $0.00027 | $0.00328 |
| Haiku 4.5 | $0.00013 | $0.00164 |
Grade A, and why
hermit-audit scanned grade A with 2 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 5d ago.
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.
Reads agent configuration directorieslowAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
cat ~/.claude/settings.json 2>/dev/null | python3 -m json.tool | grep -A5 hooks Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
Enumerates other installed skillslowAgent snooping
Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.
ls ~/.claude/skills/ 2>/dev/null | head -20 Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/hermit-audit — Hermit + CC 종합 감사
8축 평가 사이클: 구조 → 맥락 → 계획 → 실행 → 검증 → 개선 → 토큰 경제 → 모델 라우팅
체크리스트 원본: references/checklist.md (반드시 먼저 읽어 항목 정의 확인).
분석의 정의: 갖춘 것(Static)과 실제로 하는 것(Behavioral)의 gap을, User/Project 두 스코프에서 측정. 축 7-8은 Hermit 전용 — CC 환경에서는 N/A 처리.
8축 매핑 개요
| 축 | 이름 | 항목 수 | 출처 |
|---|---|---|---|
| 1 | 구조 (Structure) | 3 | check-harness 6축에서 차용 |
| 2 | 맥락 (Context) | 3 | check-harness 6축에서 차용 |
| 3 | 계획 (Plan) | 2 | check-harness 6축에서 차용 |
| 4 | 실행 (Execution) | 3 | check-harness 6축에서 차용 |
| 5 | 검증 (Verification) | 2 | check-harness 6축에서 차용 |
| 6 | 개선 (Improvement) | 2 | check-harness 6축에서 차용 |
| 7 | 토큰 경제 (Token Economy) | 6 | 신규 — Hermit 전용 |
| 8 | 모델 라우팅 (Model Routing) | 4 | 신규 — Hermit 전용 |
총 항목: 25개 (15 + 6 + 4)
축 1-6은 check-harness의 검증된 6축 프레임워크를 그대로 상속한다. 이 스킬의 핵심 가치는 축 7-8 — Hermit 환경에만 존재하는 토큰 경제 및 모델 라우팅 지표.
D-minimal 실행 절차
Python 스크립트 없이 수동 체크리스트 스캔으로 실행한다.
Step 0 — 준비
# 작업 디렉토리 생성
mkdir -p /tmp/hermit-audit/
# checklist.md 읽기 (항목 정의 확인)
cat hermit_agent/skills/hermit_audit/references/checklist.md
Step 1 — 축 1-6 (CC 하네스 공통)
각 항목을 references/checklist.md 기준으로 수동 확인.
파일 존재 여부, 설정값 유무 등 정적 지표는 bash로 검증:
# 축 1: 구조
ls ~/.claude/skills/ 2>/dev/null | head -20
ls ~/.claude/rules/ 2>/dev/null
cat ~/.claude/settings.json 2>/dev/null | python3 -m json.tool | grep -A5 hooks
# 축 2: 맥락
wc -w ~/.claude/CLAUDE.md 2>/dev/null
wc -w CLAUDE.md 2>/dev/null
# 축 4: 실행
cat ~/.claude/settings.json 2>/dev/null | python3 -m json.tool | grep -A3 mcpServers
Step 2 — 축 7 (Token Economy)
# T1: CLAUDE.md 토큰 추정 (wc -w × 1.3)
echo "Global CLAUDE.md words: $(wc -w < ~/.claude/CLAUDE.md 2>/dev/null || echo 0)"
echo "Project CLAUDE.md words: $(wc -w < CLAUDE.md 2>/dev/null || echo 0)"
# T2: rules 파일 토큰 합계
echo "Global rules tokens:"
for f in ~/.claude/rules/*.md; do
words=$(wc -w < "$f" 2>/dev/null || echo 0)
tokens=$(python3 -c "print(int($words * 1.3))")
echo " $f: ~$tokens tokens"
done
# T5: MCP 서버 수
python3 -c "
import json, os
s = json.load(open(os.path.expanduser('~/.claude/settings.json')))
mcp = s.get('mcpServers', {})
print(f'MCP servers: {len(mcp)} — {list(mcp.keys())}')
" 2>/dev/null || echo "settings.json not found or no mcpServers"
# T6: Skill phased header 채택률
grep -l "## Phase" ~/.claude/skills/**/SKILL.md 2>/dev/null | wc -l
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 5d ago First seen · 163 lines · 133 tokens per session scan A c1b2145e74f2
hermit-audit is a skill published in the GitHub repository cafitac/hermit-agent (4 stars, last pushed 9d ago), licensed MIT. It adds 133 tokens to every session and 1,640 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 2 findings (reads agent configuration directories, enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
claude_code_delegation
Delegate coding, review, and refactor tasks to Claude Code CLI through Row-Bot's approval-gated shell workflow.
task_automation
Design effective automated workflows using scheduled tasks, prompt chaining, and delivery channels.
designer_guide
Guidance for creating and editing designs using the designer tool.
row_bot_status_guide
Guidance for querying and managing Row-Bot's own configuration and logs.
design_creator
Structured workflow for creating professional designs — slide decks, documents, landing pages, app mockups, and motion storyboards.
agents_guide
Guidance for delegating focused work to child Agents.