cell-host-and-microbe
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is composed entirely of natural language instructions and lacks any executable code, shell commands, or network operations. It correctly advises users to follow official guidelines and manage data/code through standard repositories.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided manuscript data, which is an untrusted input source. However, the risk is mitigated by the complete absence of dangerous capabilities.
- Ingestion points: User manuscripts analyzed for journal fit.
- Boundary markers: None.
- Capability inventory: No file-write, network, or command execution tools are utilized.
- Sanitization: None.
Audit Metadata