travel-city-walk-by-ewg9kn
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and analyze data from external URLs linked in
references/cases.md(e.g., GoodCase.ai and X.com) to derive motion patterns and prompt structures. This exposure creates a potential for indirect prompt injection if the remote resources contain instructions intended to override agent behavior. - Ingestion points:
references/cases.mdcontains external media and source URLs. - Boundary markers: The skill lacks explicit instructions to ignore or isolate potential commands embedded within the fetched metadata or summaries.
- Capability inventory: The skill utilizes browser, image, or video inspection tools to access and process the remote content.
- Sanitization: No sanitization or validation of the external content is performed before processing.
- [EXTERNAL_DOWNLOADS]: The workflow relies on fetching reference media, summaries, and source data from
goodcase.ai,media.goodcase.ai, andx.com. These resources are used to ground the AI video prompt generation in specific evidence-based cases.
Audit Metadata