peer-review-writer

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: Analysis of the skill instructions and supporting reference documents found no evidence of malicious patterns, obfuscation, or security vulnerabilities.
  • [DATA_EXFILTRATION]: The skill includes explicit security guidelines regarding confidentiality, instructing the agent never to leak manuscripts or attempt to deanonymize authors.
  • [PROMPT_INJECTION]: The instructions contain clear usage boundaries (e.g., separating reviewer and author roles) without any patterns attempting to bypass agent safety filters or system instructions.
  • [EXTERNAL_DOWNLOADS]: All external references are limited to academic citations (arXiv IDs) used for grounding. No automated network downloads or script execution patterns were identified.
  • [COMMAND_EXECUTION]: The skill does not utilize shell commands or local file system operations beyond reading its own reference files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 02:50 PM
Security Audit — agent-trust-hub — peer-review-writer