horizontal-vertical-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions specify that the agent should ingest external user-provided materials such as documents, screenshots, links, and chat records to generate research reports. This creates a surface for potential indirect prompt injection from those external sources.\n
- Ingestion points: User-provided materials referenced in SKILL.md (documents, screenshots, links, chat records).\n
- Boundary markers: None identified in the prompt template.\n
- Capability inventory: The skill is restricted to text generation based on a fixed research methodology and does not perform autonomous file or network operations.\n
- Sanitization: No explicit sanitization or filtering of external content is described.\n- [PROMPT_INJECTION]: The skill uses authoritative language to enforce strict adherence to its research framework and prevent the agent from modifying or summarizing the core prompt. These are formatting constraints for output consistency and do not appear to be attempts to bypass safety filters.
Audit Metadata