od-video-frame-glitch-title
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill interpolates user-supplied data from the
$ARGUMENTSvariable directly into a creative brief sent to the open-design tools. This lack of sanitization or boundary markers creates a surface for indirect prompt injection, where a user could provide instructions disguised as design requirements to influence the agent's subsequent file-authoring and artifact registration actions. - Ingestion point: The
$ARGUMENTSplaceholder inSKILL.mdcaptures arbitrary user input. - Boundary markers: None. The user input is placed inside a string template without delimiters or instructions to ignore embedded commands.
- Capability inventory: The agent is instructed to use file-editing tools to author design files and call
register_open_design_artifact. - Sanitization: No validation or filtering is applied to the input before it is used in the prompt logic.
Audit Metadata