omh-content-operator
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied requests and context for content generation, creating an attack surface for indirect instructions. Evidence: 1. Ingestion points: 'user request' and 'target context' as specified in SKILL.md. 2. Boundary markers: explicit 'prepared-vs-observed' boundaries are defined to separate instructions from evidence. 3. Capability inventory: includes execution of the 'omh' command for runtime tracking. 4. Sanitization: inclusion of rules for byte-for-byte copying from canonical sources and source citation requirements.
- [COMMAND_EXECUTION]: The skill includes a platform-specific shell command to log runtime activity and status. Evidence: 'omh runtime record' command found in the Runtime Evidence section of SKILL.md.
Audit Metadata