The Agent Skills Directory

[SAFE]: The skill is a set of markdown-based instructions for security engineering. It does not contain any executable scripts (Python, Node.js, etc.) or automation that could perform malicious actions.- [PROMPT_INJECTION]: The skill includes a 'Prompt Injection Safety Notice' in Section 8 of SKILL.md. This section explicitly instructs the agent to ignore any commands embedded in analyzed content and to treat them as data only. A regex-based detector flagged a potential injection pattern (PI_IGNORE_INSTRUCTIONS), but in context, this is a defensive instruction provided to the AI agent to prevent it from following malicious commands within documents it is analyzing.- [COMMAND_EXECUTION]: Section 11 (Dynamic Context Injection) was evaluated for patterns like !command. No such execution patterns are present in the provided files. All shell-like examples are formatted as static code blocks for documentation purposes.- [DATA_EXFILTRATION]: There are no network operations or data exfiltration patterns. The instructions explicitly forbid data exfiltration in the safety notice.- [EXTERNAL_DOWNLOADS]: The skill does not perform any external downloads or package installations. It references standard security frameworks and well-known documentation from Microsoft, OWASP, and MITRE.

threat-modeling