network-security-scanner
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied requests to generate security configurations and execute bash commands. It lacks explicit instructions to sanitize inputs or implement boundary markers to isolate instructions within processed data.
- Ingestion points: User-provided network scanning patterns and requests (SKILL.md).
- Boundary markers: None identified in the skill instructions.
- Capability inventory: Broad shell command execution via Bash(cmd:*) and file modification capabilities via Write and Edit.
- Sanitization: No input validation or output escaping mechanisms are defined.
Audit Metadata