The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains defensive instructions intended to protect the agent from adversarial inputs found in reviewed data. It explicitly warns the agent to treat reviewed files as data rather than instructions and to ignore bypass markers like 'ignore previous instructions' within reviewed content. This is a security best practice for audit agents and not a malicious injection.
[DATA_EXFILTRATION]: No sensitive data access or exfiltration patterns were identified. The skill's defined scope is limited to reviewing configuration files (YAML, Terraform) for security posture.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns, package installations, or dynamic code execution methods were found. The skill operates in a read-only capacity.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not use the dynamic context injection syntax (!command).

logging