doc-cache
Transparent read-through cache for documentation lookups. Checks cache before WebFetch, caches results with expiration, garbage collects stale entries. TRIGGER when: fetching docs, API references, technical documentation, or user asks to look up library/framework usage. DO NOT TRIGGER when: user provides inline code or docs directly, or the information is already in context.
来源信息
- 仓库
- orakitine/toolbox
- 最近来源活动
- 2026年5月23日 00:42
- 检测到的 SKILL.md 语言
- 英语
- 星标
- 0
- 分支
- 0
安装方式
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
检查来源文件
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
文件资源管理器
2 个文件正在显示 SKILL.md
SKILL.md
来源说明 · 只读预览- name
- doc-cache
- description
- Transparent read-through cache for documentation lookups. Checks cache before WebFetch, caches results with expiration, garbage collects stale entries. TRIGGER when: fetching docs, API references, technical documentation, or user asks to look up library/framework usage. DO NOT TRIGGER when: user provides inline code or docs directly, or the information is already in context.
- argument-hint
- [topic | clean | list | stats]
- allowed-tools
- ["Bash","Read","Write","WebFetch","WebSearch"]
# Purpose
Transparent caching layer for documentation lookups. Before fetching docs from the web, check the cache. On miss or expiry, fetch live and cache the result. Saves tokens and time on repeated lookups. Requires `python3` in PATH (stdlib only).
## Variables
CACHE_DIR: ~/.claude/doc-cache/cache # Where cached docs live
CACHE_TOOL: python3 ./scripts/cache.py # Cache management CLI (relative to skill dir)
MAX_AGE_DAYS: 14 # Cache expiration in days
## Workflow
1. **Check Cache**
- Search for a cached doc matching the topic: `<CACHE_TOOL> find "<topic>"`
- The tool searches filenames and frontmatter (url, title, description) for matches
- IF: hit and fresh (within MAX_AGE_DAYS) → go to step 4
- IF: hit but expired → go to step 2 (re-fetch)
- IF: miss → go to step 2
- Example: `<CACHE_TOOL> find "tanstack router"` → `HIT ~/.claude/doc-cache/cache/tanstack-router-api.md (3 days old)`
- Example: `<CACHE_TOOL> find "stripe webhooks"` → `MISS`
- Tool: Bash
2. **Fetch Live**
- Use WebFetch to grab the documentation from the web
- IF: URL known → fetch directly
- IF: only topic known → WebSearch first to find the right docs page, then WebFetch
- Example: WebFetch `https://tanstack.com/router/latest/docs/framework/react/api` → markdown content
- Tool: WebFetch
3. **Cache the Result**
- Save to `<CACHE_DIR>/<slug>.md` with frontmatter:
```
---
url: <source URL>
fetched: <YYYY-MM-DD>
expires: <YYYY-MM-DD>
title: <page title>
description: <one-line summary>
---
<content>
```
- Slug derived from topic: lowercase, hyphens, no special chars (e.g. "TanStack Router API" → `tanstack-router-api`)
- IF: updating expired entry → overwrite the existing file
- Example: Save to `~/.claude/doc-cache/cache/tanstack-router-api.md`
- Tool: Write
4. **Read and Use**
- Read the cached file and use it as authoritative reference
- Cite the source: "Per [title] docs (cached YYYY-MM-DD from URL)..."
- Example: Read `~/.claude/doc-cache/cache/tanstack-router-api.md` → answer user's question
- Tool: Read
5. **Garbage Collect (On Request)**
- IF: user asks to clean up, or cache grows large
- Run: `<CACHE_TOOL> clean` — removes all expired entries
- Run: `<CACHE_TOOL> list` — shows cached docs with age and status
- Example: `<CACHE_TOOL> clean` → `Removed 3 expired entries. 8 docs remain.`
- Tool: Bash
在 GitHub 查看