The Agent Skills Directory

[SAFE]: The skill is designed to manage a specific local file (.scratch/backlog.md) for task tracking. It does not attempt to access sensitive system directories, environment variables, or credentials.\n- [SAFE]: No network operations or external dependencies were detected. The skill operates entirely within the local project environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill reads external data from the backlog file. While this data is user-controlled, the skill performs simple string manipulation (toggling checkbox markers) and does not execute or evaluate the content of the tasks, which significantly limits the risk of adversarial manipulation.

to-backlog