commercial-ad
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data, including product briefs, asset descriptions, and video content (via the
$media-analysistool). This content is used to generate scripts and creative directions which are then passed to production tools ($image-batch-runnerand$video-batch-runner). - Ingestion points: User-provided product capability facts, category details, geometry, audience descriptions, and video files processed in SKILL.md.
- Boundary markers: The skill includes output-focused constraints such as 'Claim Boundaries' and 'Generator Risks', but does not specify input delimiters or 'ignore embedded instructions' warnings for the data it processes.
- Capability inventory: The skill interacts with
$media-analysis,$image-batch-runner, and$video-batch-runnerbased on the processed user input. - Sanitization: No specific sanitization, escaping, or validation logic for user-provided strings or media metadata is implemented before the data is incorporated into the generation prompts.
Audit Metadata