The Agent Skills Directory

[SAFE]: The skill is a legitimate development tool designed to automate verification tasks. It does not exhibit any malicious behaviors such as data exfiltration, obfuscation, or unauthorized access.
[COMMAND_EXECUTION]: The skill uses standard shell commands like npm, git, and grep to analyze the local codebase. These commands are necessary for its function as a build and test verification loop.
[DATA_EXFILTRATION]: While the skill searches for sensitive strings like 'sk-' and 'api_key', it does so locally to warn the developer of potential secrets in the code. No network operations are performed to send this data externally.

verification-loop