create-video-start
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill orchestrates a pipeline where untrusted user input (the creative brief) is processed by multiple AI agents in sequence. A maliciously crafted brief could attempt to hijack the pipeline's logic to execute unintended actions.
- Ingestion points: User-provided content in the
CREATIVE_BRIEFvariable is ingested directly into the pipeline inSKILL.md. - Boundary markers: The prompts do not use delimiters (e.g., XML tags or clear separators) or specific instructions for the agent to treat the brief as data rather than instructions.
- Capability inventory: The orchestrated sub-skills are granted high-privilege tools including
Bash,Write,Edit, andGlobacross multiple files (SKILL.md,rules/claude-cli-patterns.md). - Sanitization: There is no evidence of sanitization or validation of the creative brief before it is passed to the
motion-designerskill and subsequently propagated through the pipeline.
Audit Metadata