paper-plan
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface by processing user-supplied research paper drafts. 1. Ingestion points: User-provided research paper drafts critiqued against the plan. 2. Boundary markers: No explicit delimiters are specified to wrap the untrusted draft content. 3. Capability inventory: The skill does not request or utilize any tools, shell access, or network operations. 4. Sanitization: The instructions direct the agent to report mismatches as questions rather than performing direct rewrites or executions, which serves as a mitigation against instructions contained within the draft.
Audit Metadata