nobu/research-agent-claude-code
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
Research agent for Claude Code which gives ability to process 50+ results for each request and search for domain knowledge data
108 commits 1 branch 0 tags 1.6 MiB
Find a file
Exact
Nobu 591ec851ee Fix Quick Start: add explicit CLAUDE.md copy command, mention tool routing
2026-06-10 20:36:16 +05:00
.claude Workflow updated 2026-04-27 04:12:17 +05:00
examples
.gitignore
CLAUDE.md Web research agent updated to avoid hitting rate limits 2026-04-22 21:39:17 +05:00
README.md Fix Quick Start: add explicit CLAUDE.md copy command, mention tool routing 2026-06-10 20:36:16 +05:00

README.md

Web Search Agent

Deep web search for Claude Code. Fetches and processes 50+ results per query — far beyond the typical 10-20 result limit of built-in tools.

Why use it

Most LLM search tools return 10-20 results, limiting research depth on complex questions. This tool digs deeper — more sources, smarter extraction, better answers. Works with any LLM through Claude Code.

Quick Start

git clone https://github.com/itohnobue/research-agent-claude-code
cp -R research-agent-claude-code/.claude /path/to/your/project/
cp research-agent-claude-code/CLAUDE.md /path/to/your/project/

If you already have a CLAUDE.md, append this one instead of overwriting. This teaches Claude Code to route all web searches through this tool. Test it: "Search for the most performant Rust web frameworks in 2025"

Auto-installs Python dependencies via uv on first run. No API keys required.

Usage

.claude/tools/web_search.sh "React server components best practices" --tech
.claude/tools/web_search.sh "CRISPR delivery methods" --sci --med
.claude/tools/web_search.sh "Kalman filter implementations" --sci
Flag Sources Best for
(none) DuckDuckGo + Brave + Reddit + DDG News General web
--tech + Hacker News, Stack Overflow, Dev.to, GitHub Software, DevOps
--sci + arXiv, OpenAlex CS, physics, math, engineering
--med + PubMed, Europe PMC, OpenAlex Medicine, clinical trials

Key features

  • 50+ results per query via DuckDuckGo + Brave search fallback
  • Anti-bot bypass — Scrapling with TLS fingerprinting, auto-fallback to httpx
  • Smart extraction — Trafilatura content-area detection (article body, not sidebars)
  • Token compression — sentence-level BM25 + centrality scoring, cross-page deduplication
  • Non-English support — auto-detects Japanese, Chinese, Korean, Russian, Arabic, Thai and sets appropriate DDG region
  • Wayback Machine fallback for paywalled pages
  • PDF extraction via pdftotext (poppler)

Full feature list and blocked/API-routed domain documentation in the included instructions.

Requirements

  • Python 3.11+ (auto-installed by uv if needed)
  • pdftotext (optional — from poppler, for PDF content extraction)

License

MIT