oneos-biz-rules
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, potentially untrusted data which represents an indirect prompt injection surface.
- Ingestion points: User-provided queries regarding business logic and content from sibling files like
../fayanruju/retrieval.mdare ingested into the agent's context as specified in the skill standard section. - Boundary markers: The skill lacks explicit instructions or delimiters (such as XML tags) to prevent the agent from potentially treating data within the knowledge base as instructions.
- Capability inventory: According to the skill body in SKILL.md, the agent is restricted to retrieval and adjudication tasks and is explicitly forbidden from modifying code or prototypes ('不改原型、不改真码').
- Sanitization: There are no documented procedures for validating, filtering, or sanitizing the data retrieved from external files or user input before it is processed by the model.
Audit Metadata