global-change-biology
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied manuscript data which is untrusted input.
- Ingestion points: Manuscript text provided by users for evaluation (SKILL.md).
- Boundary markers: None explicitly defined to separate manuscript content from instructions.
- Capability inventory: None. The skill does not use any tools for file writing, network access, or command execution.
- Sanitization: No sanitization is performed on user input.
- [SAFE]: No malicious patterns, obfuscation, or dangerous command executions were detected. All instructions are limited to providing editorial feedback based on journal standards. All resource references point to local documentation folders typical for this environment.
Audit Metadata