The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script (warden_agent/scan.py) found within the user's project directory. This is the intended mechanism for running the Warden security scanner and is performed within the project context.
[EXTERNAL_DOWNLOADS]: The instructions recommend installing semgrep and pip-audit via pip. These are well-known, industry-standard security tools used for static analysis and dependency auditing.
[DATA_EXFILTRATION]: The skill reads a local JSON report (.reports/warden-latest.json) to display findings to the user. No unauthorized network transmission of sensitive data was detected.

warden-scan