The Agent Skills Directory

[SAFE]: The skill provides a development environment for AI agent skills. The static analysis findings regarding subprocess.Popen and subprocess.run were reviewed and found to be benign, as they are used to execute local CLI tools (claude, lsof) and standard build processes within the developer's environment.
[SAFE]: Remote data handling in scripts/run_eval.py and scripts/improve_description.py follows standard practices for interacting with the Claude CLI, with no evidence of sensitive data exfiltration or unauthorized access.
[SAFE]: The eval-viewer component generates local HTML reports for human review of skill performance, using standard web technologies without external dependencies or hidden trackers.
[SAFE]: The packaging logic in scripts/package_skill.py correctly implements exclusion lists for build artifacts like node_modules and __pycache__, following software distribution best practices.

skill-creator