Coach AI

Instructions

Implementation Reference

• src/coach_ai/server.py:594-599 (handler)

  The MCP tool definition for 'get_user_context' which delegates to the storage layer.

  async def get_user_context() -> str:
      """Get relevant context about the user (facts, patterns, preferences).
  
      This retrieves stored facts about the user to help personalize responses.
      """
      return await storage.get_user_context()

• src/coach_ai/storage.py:356-388 (handler)

  The actual implementation of the 'get_user_context' tool logic, which queries the 'user_facts' database table and formats the retrieved data.

  async def get_user_context() -> str:
      """Get relevant context about the user.
  
      Returns:
          Formatted user context
      """
      db = await get_db()
  
      cursor = await db.execute(
          "SELECT fact, category FROM user_facts ORDER BY created_at DESC LIMIT 20"
      )
      rows = await cursor.fetchall()
  
      if not rows:
          return "No user facts stored yet. Use add_user_fact() to remember important information."
  
      result = "\n=== USER CONTEXT ===\n\n"
  
      # Group by category
      categories = {}
      for row in rows:
          cat = row["category"]
          if cat not in categories:
              categories[cat] = []
          categories[cat].append(row["fact"])
  
      for category, facts in categories.items():
          result += f"{category.upper()}:\n"
          for fact in facts:
              result += f"  - {fact}\n"
          result += "\n"
  
      return result.strip()