The Agent Skills Directory

[SAFE]: The skill's operations are confined to the local project environment, involving reading requirement files and writing plan documents. No network exfiltration or unauthorized credential access was found.
[SAFE]: Command templates provided in the planning examples (such as npm test and git commit) are standard developer practices intended for inclusion in generated plans for subsequent review, rather than immediate automated execution.
[SAFE]: The Python-based logic for DAG generation serves as a structural guideline for the model's reasoning process and does not involve the dynamic execution of untrusted code or unsafe data processing.

loop-planner