smithery.ai

Doc Vault Skill

Auto-activating documentation cache with fresh API docs. Fetches and automatically consults cached documentation when user works with libraries/frameworks.

Installation

$ npx skills add https://smithery.ai

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery.ai · top by installs.

npx skills add https://smithery.ai

Browse all from smithery.ai

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Skill metadata

Parsed from SKILL.md frontmatter.

Allowed toolsRead, WebFetch, Bash, Write
Declared agents claude-code

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 3,349 B
  • docs SUMMARY.md 178 B

History

  1. First recorded snapshot · 1 installs

SKILL.md

Purpose

Fetch and automatically consult cached documentation from official sources. Provides fresh, up-to-date API documentation that overrides stale training data. Auto-triggers on "docs" keywords, maintains session-persistent index, and always prefers cached docs over training data.

Variables

CACHEDIR: .claude/skills/doc-vault/cache # Where cached documentation files are stored INDEXFILE: .claude/skills/doc-vault/README.md # Lightweight registry of cached docs TRIGGERSENSITIVITY: conservative # Options: conservative (explicit keywords only), aggressive (broader matching) CITESOURCES: true # Always cite source URL and date when using cached docs INDEXTOOL: .claude/skills/doc-vault/tools/manageindex.py # Tool for managing the doc index

Workflow

  1. Detect Trigger

- Check for "docs" or "documentation" keywords (conservative mode) - Tool: Pattern matching on user input - Example: "check the TanStack Query docs" → TRIGGERED

  1. Load Index (First Trigger Only)

- IF first trigger in session, READ INDEXFILE - Tool: Read .claude/skills/doc-vault/README.md - Loads into context, persists for entire session - Example: INDEXFILE loaded → 3 cached docs available (TanStack Query, React, Zod)

  1. Route to Cookbook

- Analyze user request to determine action type - Options: fetch new doc, consult cached doc, list available docs - Example: "save docs from URL" → Fetch scenario, "check the docs" → Consult scenario

  1. Execute Cookbook Scenario

- Read and follow the appropriate cookbook file - Tool: Read cookbook markdown and execute instructions - Example: Consult scenario → Read cached doc → Provide answer with citation

  1. Cite Sources

- IF using cached docs and CITE_SOURCES is true - Always state source URL and date - Example: "According to TanStack Query docs (cached 2025-12-11 from https://tanstack.com/query/latest)...";

Cookbook

Fetch New Documentation

  • IF: User provides URL with "save docs from" or "cache docs"
  • THEN: Read and execute .claude/skills/doc-vault/cookbook/fetch-new-doc.md
  • EXAMPLES:

- "save docs from https://zod.dev as zod-validation" - "cache docs: https://prisma.io/docs/orm as prisma-orm" - "add to doc vault: https://trpc.io/docs as trpc-api"

Consult Cached Documentation

  • IF: User mentions "docs" OR "documentation" OR "api reference" (conservative trigger)
  • THEN: Read and execute .claude/skills/doc-vault/cookbook/consult-cached.md
  • EXAMPLES:

- "check the TanStack Query docs and implement caching" - "according to the latest Prisma docs, add migrations" - "consult the React docs for Suspense" - "use fresh Zod docs for validation"

List Available Documentation

  • IF: User asks what documentation is cached
  • THEN: Read and execute .claude/skills/doc-vault/cookbook/list-docs.md
  • EXAMPLES:

- "what docs do we have?" - "list cached documentation" - "show me the doc vault" - "what's in the cache?"