clay-animation-video-prompt
Warn
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: MEDIUMPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill incorporates a "PORTFOLIO/SPEC MODE" that explicitly commands the agent to avoid sanitizing or rejecting requests involving medical claims or fictional product benefits. This instruction is intended to circumvent safety guardrails that prevent the generation of misleading or unethical advertising content.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts external product briefs and ad scripts for processing without implementing boundary markers or input sanitization. This establishes a vulnerability surface where instructions hidden within user-supplied data could alter the agent's behavior. • Ingestion points: User-provided product briefs, scripts, and reference descriptions in SKILL.md. • Boundary markers: Absent. • Capability inventory: The skill is restricted to text generation and does not have file-writing or network capabilities. • Sanitization: There is no process for filtering or escaping user-supplied text before its inclusion in prompts.
Audit Metadata