talking-head-and-piece-to-camera
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process scripts generated by other skills (e.g.,
short-form-video-script). While this creates a potential surface for indirect prompt injection where an attacker could influence the coaching output, the risk is negligible as the skill lacks any dangerous capabilities such as file system writes, network requests, or command execution. - Ingestion points: Scripts from
short-form-video-scriptandyoutube-long-form. - Boundary markers: The instructions explicitly scope the agent's role to coaching and formatting (beat maps/prompter text).
- Capability inventory: No subprocess calls, network operations, or file-writing capabilities are present.
- Sanitization: Not applicable as the output is purely instructional text for a human.
- [SAFE]: The skill references several well-known external services for video editing and avatar generation (e.g., CapCut, Descript, Heygen, Synthesia). These references are documented neutrally as legitimate parts of the video production workflow.
- [SAFE]: The skill coordinates with the vendor-specific platform
WoopSocialfor the publishing stage of the workflow, which is consistent with its stated purpose.
Audit Metadata