marketing-video
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided audio and video recordings to generate clips and summary cards, creating a surface for potential indirect prompt injection.
- Ingestion points: Workflow in REPURPOSING.md takes long-form recordings as input for automated processing.
- Boundary markers: No specific delimiters or safety instructions are defined for the transcript data processed by the models.
- Capability inventory: The skill interacts with video editing platforms and generation APIs (HeyGen, Veo).
- Sanitization: Verification involves human review checkpoints, such as inspecting the first few generated videos of a batch.
- [COMMAND_EXECUTION]: The skill uses ffprobe for technical verification of video assets.
- Evidence: Mentioned in the 'Prove it' sections of SKILL.md, AI-GENERATION.md, and REPURPOSING.md to verify file properties like resolution and duration.
- [EXTERNAL_DOWNLOADS]: The skill references documentation and services from trusted and well-known providers.
- Evidence: Links to official technical and safety documentation from Google, OpenAI, Adobe, and the W3C.
Audit Metadata