derive-first-principles
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill evaluates external context to identify underlying principles.
- Ingestion points: Reads and analyzes the user's request and task context as specified in
SKILL.md. - Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential injection attempts within the analyzed context.
- Capability inventory: No sensitive capabilities such as network access, file system writes, or subprocess execution are defined in this skill.
- Sanitization: Input context is processed without specific sanitization or filtering logic.
- [SAFE]: The skill uses a GitHub repository for installation which is consistent with the author's identity.
Audit Metadata