project-focus-first
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill uses strong steering language such as 'Non-Negotiable Rule', 'Hard gate', and 'Forbidden Behavior' to control agent behavior. However, these are legitimate workflow constraints rather than attempts to bypass safety guidelines. The skill explicitly states it does not override system instructions or safety policies.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user input to define project categories and manage internal conversation state. While this represents a data ingestion surface, the skill lacks high-risk capabilities (such as network access or file system modifications) that could be exploited via injection. It also includes logic for handling ambiguous or invalid user responses.
- [SAFE]: The skill is a standard procedural guideline (Clarification Gate) for project management. It contains no executable scripts, external dependencies, or network-bound operations. All referenced files are local documentation or evaluation data.
Audit Metadata