MCP server for co-lleague — AGI co-worker (screen, predict, intent)

[Malek-Ghorbel]

Summary

Adds the co-lleague MCP server — an AGI co-worker that provides screen analysis, tabular prediction, intent execution, and agent status monitoring.

Tools

• co-lleague_screen_query — Query screen content with AI analysis
• co-lleague_tabular_predict — Run tabular predictions on feature arrays
• co-lleague_intent_execute — Execute natural-language intents
• co-lleague_agent_status — Check agent online/offline state and activity

Package

• @modelcontextprotocol/server-co-lleague
• Wraps the @aimino/co-lleague-mcp npm package
• MIT License

Directory Structure

src/co-lleague/
  README.md       — Installation, configuration, examples
  index.ts        — MCP server with 4 tools + resource templates
  package.json    — npm package config
  tsconfig.json   — TypeScript configuration
  vitest.config.ts— Test configuration
  __tests__/      — Unit tests

Configuration

Users set CO_LEAGUE_API_KEY in env to authenticate with the co-lleague backend.