web-researcher

Web Researcher Plugin

Iterative web research agent that searches, discovers new directions, and synthesizes findings - like a human researcher would.

Installation

claude plugin marketplace add fprochazka/claude-code-plugins --scope user
claude plugin install web-researcher@fprochazka-claude-code-plugins --scope user

Usage

/web-researcher:search <your query>

Example:

/web-researcher:search how to implement OAuth2 PKCE flow in a CLI application

How It Works

The plugin spawns a dedicated research subagent that:

1. Searches with your initial keywords
2. Reads 2-3 promising pages to gain context
3. Discovers new terminology, related concepts, and authoritative sources
4. Follows leads by searching with newly discovered keywords
5. Repeats 2-4 iterations until the query is comprehensively answered

The subagent is restricted to only WebSearch and WebFetch tools, and all exploratory dead ends stay isolated in the subagent context - only the distilled, useful findings are returned.

Output Format

Results are structured as:

• Summary - comprehensive answer to your query
• Key Findings - most important points discovered
• Sources - only URLs that were actually helpful

Components

• /web-researcher:search - Command to trigger research
• web-researcher agent - The iterative research subagent

License

MIT