academic-book-reviewer
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: No malicious patterns were identified. The skill is composed entirely of natural language instructions and configuration metadata for academic review tasks.
- [PROMPT_INJECTION]: The skill processes untrusted external data (manuscripts and outlines), which presents a surface for indirect prompt injection where a manuscript could contain instructions to influence the agent's behavior.
- Ingestion points: Manuscript, outline, or chapter set inputs defined in the Agent Contract within
SKILL.md. - Boundary markers: None; the instructions do not specify delimiters to isolate untrusted manuscript content from the agent's task instructions.
- Capability inventory: None. The skill does not have access to tools, network operations, or file system write capabilities, significantly limiting the impact of any potential injection.
- Sanitization: Not present; the manuscript content is analyzed as-is for the purpose of generating a review.
Audit Metadata