The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill focuses on project management tasks like creating and organizing issues using a specialized CLI tool.
[COMMAND_EXECUTION]: The skill uses a command-line utility called bd to interact with an issue tracking system. These operations (create, show, dep add) are confined to task management and do not involve system-level modifications or arbitrary shell access.
[DATA_EXFILTRATION]: While the skill reads the codebase to understand context, there are no network operations, external URLs, or data transmission patterns identified that would indicate exfiltration risks.
[REMOTE_CODE_EXECUTION]: The skill includes instructions to spawn a secondary 'Plan Reviewer' agent using a local skill path (.claude/skills/reviewer-plan/SKILL.md). This is a standard internal delegation mechanism and does not involve remote code or untrusted scripts.

planner