peer-review

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • Indirect Prompt Injection (LOW): The skill processes untrusted external manuscripts, creating a surface for indirect prompt injection.\n
  • Ingestion points: Employs Read for local file access and WebFetch for remote content retrieval.\n
  • Boundary markers: Absent. The protocol does not define delimiters or specific instructions to ignore embedded commands within processed manuscripts.\n
  • Capability inventory: Includes WebSearch, WebFetch, Read, Grep, and Glob, enabling both file system interaction and outbound network communication.\n
  • Sanitization: Absent. There is no evidence of input validation or content sanitization before the data is processed by the agent.\n- Data Exposure & Exfiltration (LOW): The toolset includes WebSearch and WebFetch, which communicate with external domains. While these are used for research and verification, they constitute an outbound network channel to non-whitelisted domains. No sensitive file access or credential exposure was detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:28 PM