notis-reports
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes and renders interactive report views and HTML pages which could contain instructions intended to influence agent behavior.
- Ingestion points: Content is ingested via the LOCAL_NOTIS_RENDER_VIEW tool and through sibling references to the notis-apps skill.
- Boundary markers: The instructions do not specify markers or warnings to ignore instructions embedded within the processed reports.
- Capability inventory: The skill can save to Notis Notes, update existing notes, and generate public share links upon request.
- Sanitization: No explicit sanitization or escaping of rendered content is mentioned in the instructions.
- [SAFE]: The skill includes strong anti-exfiltration instructions, explicitly prohibiting the saving of information to external databases, websites, paste sites, or public file hosts, ensuring data remains within the intended platform.
Audit Metadata