The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a Tool Execution Protocol in SKILL.md that uses standard shell commands (cat, mkdir, stat) to write and validate report files on the local filesystem. These operations are essential for its function as a reporting tool and are used safely.
[PROMPT_INJECTION]: The skill is designed to process untrusted security data (findings, evidence), which constitutes a surface for indirect prompt injection. Ingestion occurs from verified findings and recon data; while the skill lacks automated boundary markers or sanitization for output markdown, it provides manual redaction guidelines for PII.
[SAFE]: The skill is entirely focused on documentation methodology. It contains no obfuscated code, does not perform unauthorized network operations, and includes no persistence mechanisms or hardcoded credentials. It properly references standard security libraries like the Python cvss package for scoring.

writing-security-reports