animation-style-consistency-by-1wr1qau
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references multiple external URLs hosted on
goodcase.aiandx.com. These links point to video files (.mp4), image posters (.jpg), and social media posts intended to serve as visual references for the agent to analyze using its tools. - [INDIRECT_PROMPT_INJECTION]: The skill processes external 'Case evidence' which includes prompt excerpts and summaries from third-party sources to generate its own output.
- Ingestion points: The
references/cases.mdfile contains structured data, descriptions, and URLs that the agent is instructed to read and inspect. - Boundary markers: The skill includes a 'Safety and attribution' section and a 'Verification' checklist that requires the agent to distinguish between observed evidence and its own recommendations.
- Capability inventory: The workflow explicitly requests the use of image, video, or browser tools to inspect external media at the provided URLs.
- Sanitization: There is no programmatic sanitization of the input text; the skill relies on instructional constraints to prevent identity drift or theme contamination.
Audit Metadata