---

name: todo-manager description: Manage Todo operations (Create, Read, Update, Delete) using SQLModel and Postgres. Use when the user wants to organize tasks.

Todo Management Skill

Executive Summary

This skill governs the lifecycle of Todos in the "Advanced AI-Todo" application. It ensures all operations are persistent, async-ready, and compliant with the project's technical mandates.

Operational Mandates

1. Environment First: ALWAYS run source .venv/bin/activate before any backend execution or uv command.
2. Dependency Management: Use uv for all package operations. Never use raw pip.
3. Connectivity: Connect to Gemini 2.5 Flash using the AsyncOpenAI class.
4. Docs Validation: Before proposing code changes to SQLModel or FastAPI, call the context-7 MCP server to verify the latest async patterns.

Execution Workflow

• Step 1: Planning: Check SKILLS/SKILL.md for current project state.
• Step 2: Documentation: Query MCP context-7 for "FastAPI SQLModel Postgres async best practices".
• Step 3: Implementation:
  • Use AsyncOpenAI for the chatbot logic.
  • Ensure Postgres connection strings are pulled from environment variables.
  • Implement CRUD operations as async functions.
• Step 4: Verification: Run tests using uv run pytest.

Examples

• Input: "Create a task to buy milk."
• Action: Invoke todo-manager → Search MCP for SQLModel patterns → Generate async code to insert into Postgres → Confirm with Gemini 2.5 Flash.