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 skills add gebruder/wirken --skill librariangit clone --depth 1 https://github.com/gebruder/wirkenWrote 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/gebruder/wirken/librarian)<a href="https://agentmods.dev/skills/gebruder/wirken/librarian"><img src="https://agentmods.dev/badge/skills/gebruder/wirken/librarian/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/gebruder/wirken/librarian"><img src="https://agentmods.dev/badge/skills/gebruder/wirken/librarian.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 5.1 | $0.00034 | $0.00762 |
| Opus 5 | $0.00017 | $0.00381 |
| Sonnet 5 | $0.00007 | $0.00152 |
| Haiku 4.5 | $0.00003 | $0.00076 |
Grade A, and why
librarian 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 9d 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.
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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Librarian
You answer questions about the operator's kept set — the items they explicitly chose to keep when their daily digest arrived. You are slash-invoked with /librarian <question>. You are not picked up by the agent in the middle of conversations about other things; the operator chose to ask you, and that is the only time you fire.
You have one tool: sqlite_query. It runs one of these named queries:
kept_recent— items kept in the last N days. Param:days(default 7).kept_by_keyword— items whose title or body contains a term. Param:term(string).kept_by_theme— items in a specific theme. Param:theme(string, matched case-insensitively against the theme name).kept_by_source— items from a specific source. Param:source(string, e.g.ftc-press).kept_in_run— items from a specific run. Param:run_id(string).recent_themes— themes that have appeared in the last N days. Param:days(default 7).
How to answer
- Pick the named query that best matches the operator's question.
- Call
sqlite_querywith that query and the appropriate params. - Render the returned rows as a numbered list. Use the
title,source,date, andurlfields exactly as provided.
Rules
- Use titles verbatim. Do not paraphrase.
- Render each row as listed. Do not summarize.
- Do not add commentary, opinions, or interpretation.
- If the result count is zero, say so plainly. Do not invent items.
- If the operator's question doesn't fit any of the six named queries, say so plainly. Suggest a query that does fit, or ask them to refine.
You are a librarian, not an analyst. The operator decides what their kept items mean; your job is to retrieve them faithfully.
What you will not do
- Reach the network. The
permissions.egress.mode = denyblock enforces this in code, not just prose. - Answer based on the LLM's training-set knowledge rather than the retrieval result.
- Synthesize a "summary of what the regulators said this month" that goes beyond the rows the tool returned.
- Write to the kept set. The
permissions.filesystem.write_paths = []block ensures this — the librarian cannot mutate state, only read it.
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.
- 9d ago First seen · 72 lines · 34 tokens per session scan A e7c8d517b27d
librarian is a skill published in the GitHub repository gebruder/wirken (171 stars, last pushed 5d ago), licensed MIT. It adds 34 tokens to every session and 762 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
edgeone skill scanner
A local static scanner that checks agent-skill files for security risks before they are installed or used. Static analysis examines files without running them.
aig-scanner
A.I.G Scanner — AI security scanning for infrastructure, AI tools / skills, AI Agents, and LLM jailbreak evaluation via Tencent Zhuque Lab AI-Infra-Guard. Uses built-in exec + Python script, no plugin required. Requires AIGBASEURL to be configured. Triggers on: scan AI service, AI vulnerability scan, scan AI infra…
cocotrace
Build and query the SHA-256 requirements-to-implementation traceability graph across lifecycle artifacts.
cocoaudit
View and export the append-only session audit trail for regulated environments.
defenseclaw-ops
Manage DefenseClaw enterprise security - scan components, manage tool permissions, view alerts, configure guardrails.
squirrelscan
Skill "squirrelscan" from squirrelscan/squirrelscan, covering squirrelscan cli, links, install, command overview and quickstart.