Docs Seeker

Execute scripts to fetch technical documentation from llms.txt sources (context7.com) with automatic query classification and agent distribution strategy.

When to Use

Need topic-specific docs (features/components/API methods)
Looking up library/framework documentation fast
Analyzing GitHub repositories for architecture
Large doc sets requiring parallel agent strategy

Key Capabilities

Capability	What It Does
Query Detection	Auto-classifies topic-specific vs general queries
Smart Fetching	Constructs context7 URLs, handles fallback chains
Result Analysis	Categorizes URLs, recommends 1/3/7 agent strategies
Zero-Token Scripts	All logic in Node.js scripts, no prompt overhead

Common Use Cases

Topic-Specific Lookup

Who: Developer needing specific feature documentation Prompt: “How do I use date picker in shadcn?”

node scripts/detect-topic.js "<query>"  # → {topic, library, isTopicSpecific}
node scripts/fetch-docs.js "<query>"    # → 2-3 focused URLs
# Read with WebFetch

General Library Docs

Who: Developer exploring new framework Prompt: “Get Next.js documentation”

node scripts/detect-topic.js "<query>"              # → {isTopicSpecific: false}
node scripts/fetch-docs.js "<query>"                # → 8+ URLs
cat llms.txt | node scripts/analyze-llms-txt.js -   # → Agent strategy
# Deploy parallel agents per recommendation

Repository Analysis

Who: Team lead investigating library architecture Prompt: “Analyze shadcn/ui repository structure”

node scripts/fetch-docs.js "github.com/shadcn/ui"   # → Repo docs
# Read with WebFetch for architecture insights

Multi-Agent Documentation Research

Who: Tech lead needing comprehensive framework knowledge Prompt: “Research React Server Components in Next.js 15”

node scripts/fetch-docs.js "<query>"                # → Multiple URLs
cat llms.txt | node scripts/analyze-llms-txt.js -   # → "7 agents recommended"
# Spawn parallel research agents

Quick Reference

Three-Script Workflow:

# 1. Detect query type
node scripts/detect-topic.js "<query>"

# 2. Fetch documentation
node scripts/fetch-docs.js "<query>"

# 3. Analyze for agent strategy (if 8+ URLs)
cat llms.txt | node scripts/analyze-llms-txt.js -

Agent Distribution:

1 agent: Few URLs (3-5)
3 agents: Medium coverage (6-12)
7 agents: Comprehensive (13+)
Phased: Massive doc sets (30+)

Environment Config:

process.env > .claude/skills/docs-seeker/.env > .claude/skills/.env > .claude/.env

Pro Tips

Scripts handle all URL construction and fallback logic automatically
Topic queries return 2-3 focused URLs (10-15s), general queries return 8+ (30-60s)
Use analyze-llms-txt.js to get parallel agent recommendations for large doc sets
Scripts are zero-token execution - no context loading overhead
Not activating? Say: “Use docs-seeker skill to fetch documentation for [library/topic]“

Research - Documentation research workflows
Planning - Plan with documentation context
MCP Management - Manage MCP servers for extended capabilities

Key Takeaway

Script-first documentation discovery with automatic query classification, intelligent URL fetching via context7.com, and parallel agent distribution for comprehensive doc coverage.