Tencent/BrowserSkill: You.com web search CLI integration

[mouse-value-add]

Source Repository

https://github.com/Tencent/BrowserSkill

Selection Rationale

Selected Tencent/BrowserSkill (1,123⭐) as the best candidate for You.com integration due to:

Perfect Integration Fit:

• Browser automation platform designed for AI agents (Cursor, Claude Code, Codex, etc.)
• Agents controlling browsers desperately need web search context and research capabilities
• CLI-based architecture makes adding bsk search command natural and non-invasive
• Already supports multiple agent frameworks - web search enhances all of them

Active Maintenance:

• Very recent commits (pushed 2026-08-18T06:10:21Z - today!)
• Active development with comprehensive documentation and examples
• MIT licensed, well-maintained TypeScript/Rust codebase
• 1,123 stars indicating strong adoption

Technical Excellence:

• Rust CLI with TypeScript browser extension - professional architecture
• Clear command structure in crates/bsk-cli/src/cli/mod.rs
• Existing pattern for adding new commands to the CLI
• Strong testing infrastructure and proper error handling

Integration Approach:

• Add bsk search <query> command following existing CLI patterns
• Optional integration via YDC_API_KEY environment variable
• Returns structured JSON results that agents can process
• Works alongside existing browser automation commands

Implementation Details

Integration Type: CLI Command Extension
Planned Environment Variables:

• YDC_API_KEY - Optional You.com API key for authenticated access
• YDC_BASE_URL - Optional API base URL override (default: https://api.you.com)

Expected Usage Example:

# Search for information agents can use to inform their browser actions
bsk search "Next.js deployment best practices" --json
bsk search "React testing patterns" --count 10

Validation Plan:

• Rust compilation and CLI integration testing
• You.com API integration with authenticated and keyless modes
• JSON output formatting compatible with agent parsing
• Error handling and fallback behavior testing
• No breaking changes to existing browser automation functionality

Candidate Evaluation Summary

Evaluated candidates:

1. Tencent/BrowserSkill (1123 stars) - SELECTED - Browser automation with perfect search integration fit
2. dcouple/Pane (389 stars) - AI agent manager, good but less specific integration point
3. DAWNCR0W/affine-mcp-server (257 stars) - MCP server, good fit but smaller impact

Selection rationale: BrowserSkill has the most natural integration fit - browser automation agents need web search context to make informed decisions about what to do on web pages. The CLI architecture makes integration clean and the active maintenance suggests good PR acceptance likelihood.

Links

• Source Repository: https://github.com/Tencent/BrowserSkill
• Branch: (TBD)
• PR: (TBD)

[mouse-value-add]

Implementation Complete ✅

Source Repository: https://github.com/Tencent/BrowserSkill
Integration Branch: https://github.com/mouse-value-add/BrowserSkill/tree/feat/youcom-search-integration
Pull Request: Tencent/BrowserSkill#106

Implementation Summary

Successfully integrated You.com search as a new CLI command in BrowserSkill:

• Integration Type: CLI command (bsk search)
• Environment Variables:
  • YDC_API_KEY (optional) - You.com API key for authenticated requests
  • YDC_BASE_URL (optional) - Custom API endpoint (defaults to https://api.you.com)
• Setup: Works immediately without configuration, optional API key enhances capabilities

Usage Examples

# Basic web search  
bsk search "React testing patterns"

# Agent-friendly JSON output
bsk search "JavaScript performance" --json --count 5

# Integration with browser automation
DOCS_URL=$(bsk search "Next.js docs" --json --count 1 | jq -r '.results[0].url')
bsk navigate "$DOCS_URL"

Validation Performed

✅ Build: Clean compilation with cargo build --release
✅ CLI Integration: bsk search --help shows proper command structure
✅ Search Functionality: Successfully queries You.com API
✅ JSON Output: Machine-readable format for agent processing
✅ Error Handling: Graceful failure modes and clear error messages
✅ No Breaking Changes: Existing BrowserSkill functionality unchanged

Fallback Behavior

• Without API key: Uses You.com's public search endpoints
• Network errors: Clear error messages with retry guidance
• Invalid parameters: Input validation with helpful corrections
• API failures: HTTP error details for debugging

Integration Benefits

This integration provides AI agents with research capabilities that complement BrowserSkill's browser automation:

1. Pre-navigation Research: Find relevant URLs before visiting sites
2. Context Gathering: Research technologies/concepts before page interactions
3. Dynamic Workflows: Use search results to inform navigation decisions
4. Knowledge Enhancement: Access real-time information for agent decision-making

The implementation follows BrowserSkill's existing patterns and maintains backwards compatibility while adding powerful new capabilities for AI agent workflows.