formula-derivation
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external theory notes and formula drafts to produce derivation packages, creating a surface for indirect instruction injection.
- Ingestion points: The workflow reads user-specified files, referenced drafts, and local research notes as defined in SKILL.md Step 1.
- Boundary markers: While the skill requires the agent to state its interpretation before manipulation, it does not utilize explicit structural delimiters to separate untrusted data from instructions.
- Capability inventory: The agent is authorized to use the Write and Edit tools to create or modify local project files like DERIVATION_PACKAGE.md (Step 7).
- Sanitization: There is no evidence of sanitization, filtering, or escaping mechanisms for the content read from external files before processing.
Audit Metadata