The Agent Skills Directory

[SAFE]: The skill is composed of instructional markdown and configuration files designed to guide the agent through a threat modeling process. It does not include any executable scripts or binary files.
[SAFE]: Instructions within prompts/threat-modeling.md and SKILL.md strictly prohibit high-risk activities such as calling real APIs, logging into systems, or reading actual credentials, which mitigates data exfiltration and unauthorized access risks.
[SAFE]: The skill implements a 'human-in-the-loop' requirement for critical security decisions, explicitly stating that risk acceptance, exception authorization, and release judgments must be left to a human user.
[SAFE]: No obfuscation, prompt injection attempts, or malicious persistence mechanisms were detected across the analyzed files.

threat-modeling