research-intelligence
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- PROMPT_INJECTION (SAFE): No malicious instructions or bypass attempts detected in the skill patterns.
- INDIRECT_PROMPT_INJECTION (SAFE): The skill processes untrusted research content but has no side-effect capabilities. 1. Ingestion points: User-provided research content (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: None (no scripts). 4. Sanitization: Absent.
- NO_CODE (SAFE): The skill consists entirely of markdown files and does not include any scripts or binary executables.
Audit Metadata