The Agent Skills Directory

[SAFE]: The skill's logic is entirely internal to the agent's workflow for project management. It reads local specification files (docs/specs/*/spec.md) and manages task tracking files (docs/specs/*/tasks/*.md) based on a provided template.
[SAFE]: The execution logic focuses on organizational structure and documentation consistency. It specifically prohibits predicting code structures (files, functions) until the implementation phase, minimizing the risk of hallucinated or insecure code generation at the task-definition stage.
[SAFE]: It handles task state transitions (pending, in-progress, completed, blocked, superseded) safely, preserving historical evidence of work without deleting or archiving files in a way that would obscure the project's security audit trail.

split-into-tasks