The Agent Skills Directory

[SAFE]: The skill implements a 'Sensitive-Files Guardrail' that explicitly prevents the agent from reading files containing secrets, credentials, or private keys, specifically blacklisting patterns like .env, .pem, .key, and .aws/credentials.
[SAFE]: A 'Post-exploration secret scan' is included in the final phase to identify and redact common secret patterns (API keys, passwords, tokens) from the output before it is presented to the user.
[COMMAND_EXECUTION]: The skill uses the 'Bash' tool to perform standard repository analysis tasks such as directory listing ('ls'), file searching ('find'), and version control history ('git log'). These commands are used as intended for codebase exploration and do not involve untrusted input or shell injection vectors.
[DATA_EXFILTRATION]: The skill is designed for local observation and reporting. No network-based exfiltration patterns or unauthorized remote connections were detected.

codebase-overview