domain
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process domain-specific documentation, instructions, and contracts from the host repository to clarify vocabulary. It also includes capabilities to update these contracts. Ingestion points: SKILL.md procedure for reading domain documentation, repository instructions, and contracts. Boundary markers: The skill contains explicit instructions to cite sources and includes a 'synonym smuggling' section designed to prevent semantic drift and unauthorized authority changes in domain terms. Capability inventory: The skill is authorized to update existing domain contracts (update_existing_domain_contracts). Sanitization: Employs semantic distinctions (Given/When/Then) and manual refinement rather than automated sanitization.
- [EXTERNAL_DOWNLOADS]: The skill provides a reference link to external engineering guidelines hosted on a public GitHub repository for domain modeling best practices.
- [SAFE]: The provided scripts are local validators for the skill's own file structure and presence of required contract files. They do not perform network operations or execute untrusted code.
Audit Metadata