Task Execution Engine

What it does

The Task Execution Engine automates implementation tasks directly from design documents formatted with markdown checkboxes. It manages these tasks, allowing an AI agent to break down complex projects into smaller, executable steps and track progress without needing separate session files. The engine uses a command-line interface (CLI) for managing task status updates like marking tasks as complete or failed.

When to use it

• Executing implementation tasks outlined in design documents.
• Managing development workflows where tasks are defined with markdown checkboxes.
• Automating the process of breaking down large projects into smaller, manageable steps.

Key capabilities

• Task management via markdown checkboxes within a design document.
• CLI commands for next, done, fail, and status task updates.
• Support for defining task priority, phase, dependencies, and associated files.
• Automated execution loop to process tasks until completion.

Example prompts

• "Run the next implementation task from my design document: design.md."
• "Mark the 'Create User model' task as completed in design.md."
• "Show me the status of all tasks in design.md."

Tips & gotchas

• Requires a design document formatted according to the specified markdown task format (see references/task-format.md for details).
• The engine operates in "unattended mode," meaning it will not stop for questions or clarification and makes autonomous decisions based on codebase patterns.
• If a task cannot be completed, it should be marked as failed with a reason.