The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown-based instructional playbooks and configuration references. No malicious code, obfuscation, or unauthorized access patterns were detected across any of the analyzed files.
[COMMAND_EXECUTION]: Maintenance playbooks, such as worktree-cleanup.md, contain instructions for using standard shell and git commands to manage development environments. These workflows include safety boundaries that require the agent to seek user confirmation for uncommitted work and verify system state before proceeding with deletions.
[PROMPT_INJECTION]: The skill includes explicit security guidelines to mitigate indirect prompt injection. For example, the babysit.md playbook instructs the agent to treat all external review text as untrusted evidence rather than executable instructions, establishing a strong defense against data-driven attacks.
[EXTERNAL_DOWNLOADS]: No unverified or suspicious external downloads were found. References to external URLs are limited to upstream attribution and documentation on well-known development platforms like GitHub.

pstack