deep-reading-analyst
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content provided by users, which could contain malicious instructions designed to subvert the agent's behavior.\n
- Ingestion points: Untrusted data enters the context via user-provided URLs and long-form content as specified in
SKILL.md.\n - Boundary markers: The instructions do not define clear delimiters or warnings to separate analyzed content from the agent's core logic.\n
- Capability inventory: The skill utilizes the
web_searchtool. It does not demonstrate capabilities for file system access, arbitrary code execution, or network exfiltration.\n - Sanitization: No validation or sanitization mechanisms are employed to process the external data safely.
Audit Metadata