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/sdeenadmin/skill-valet/skillnpx skills add SDeenAdmin/skill-valet --skill skillgit clone --depth 1 https://github.com/SDeenAdmin/skill-valetWhat 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 | $0.00091 | $0.00638 |
| Opus 5 | $0.00046 | $0.00319 |
| Sonnet 5 | $0.00018 | $0.00128 |
| Haiku 4.5 | $0.00009 | $0.00064 |
Grade A, and why
skill-valet 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 2d 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 — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.
skill-valet
skill-valet is a UserPromptSubmit hook that keeps the per-turn skill list
lean. It keyword-classifies each prompt into skill groups (defined in
~/.claude/skill-valet.json), moves a group's skill directories back into
~/.claude/skills/ when the prompt needs them, and parks groups that have
been idle for park_after turns (default 20) into ~/.claude/skills-off/.
When invoked, do this
- Run
python3 <path-to>/skill_valet.py statusto show the user which groups are ON / PARKED / PARTIAL, the turn counter, and idle times. - If the user asks why a skill is missing: check whether it belongs to a
parked group (status output) or sits in
~/.claude/skills-off/. - Manual overrides on request:
on <group>/off <group>— force a group in or out nowall-on— restore every managed skillpause/resume— kill switch: while paused the hook does nothing
- To change what is managed: edit
~/.claude/skill-valet.json— move a skill name betweenalways_on(never touched) and a group'sskillslist, or tune the group'skeywords.
Honest boundaries — state these when relevant
- Next-turn effect, not same-turn. A skill un-parked by this prompt
appears in the skill list from the NEXT turn onward. If the user needs a
parked skill right now, run
on <group>and ask them to re-send the request. - User skills only. Plugins, plugin skills, and MCP servers are bound to session start — skill-valet cannot switch those mid-session; a restart (or profile switcher) is required.
- Hooks are NEVER touched — by design. Hooks are the user's deterministic safety layer; skill-valet has no code path that reads or modifies them.
- Multi-session caveat. Directory moves are global: parking a group affects every concurrently running Claude Code session on this machine.
- Fail-open. If the hook errors it silently does nothing (exit 0) — it
will never block a prompt. If behavior seems off, check
~/.claude/skill-valet-state.jsonand the config for JSON validity.
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.
- 2d ago First seen · 45 lines · 91 tokens per session scan A 0999fcf3195a
skill-valet is a skill published in the GitHub repository SDeenAdmin/skill-valet (6 stars, last pushed 1mo ago), licensed MIT. It adds 91 tokens to every session and 638 once invoked, about $0.0005 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 skills, from other repositories
init-libra
Scaffolds the Libra docs structure for a project by reading the codebase and asking the developer about vision and domain, then writing docs plus root AGENTS.md and ARCHITECTURE.md directly to disk. Use when the user runs /init-libra or asks to scaffold Libra docs, init project docs, or set up docs structure.
update-libra
Syncs Libra-style docs (docs/decisions, docs/specs, docs/design, docs/plans) after meaningful conversations. Use when the user asks to sync docs, or at end of session when conversation produced decisions, constraints, scope changes, or progress worth recording.
flutter-app
Bootstrap a new Flutter mobile app with clean architecture, Riverpod, FVM-pinned SDK, current packages, and no deprecated APIs. Use when the user wants to start, scaffold, or set up a new Flutter app, a cross-platform mobile app, an Android or iOS app in Dart, or asks to "create a new flutter app". Handles BYOK LLM…
kn-research
Use when you need to understand existing code, find patterns, search project knowledge, investigate current external facts, or explore a large codebase before implementation.
api-testing
HTTP API testing for TypeScript (Supertest) and Python (httpx, pytest). Test REST APIs, GraphQL, request/response validation, authentication, and error handling.
api-gateway-configuration
Configures API gateways for routing, authentication, rate limiting, and request transformation in microservice architectures. Use when setting up Kong, Nginx, AWS API Gateway, or Traefik for centralized API management.