od-video-frame-light-leak-cinema
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied input via the
$ARGUMENTSvariable and incorporates it into a brief for theprepare_open_design_brieftool. - Ingestion points: User input is ingested through the
$ARGUMENTSplaceholder in theSKILL.mdfile. - Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the
$ARGUMENTSstring. - Capability inventory: The agent is instructed to call several MCP tools (
prepare_open_design_brief,list_open_design_design_systems,recommend_open_design_design_systems,register_open_design_artifact) and use its own file-editing tools to author content. - Sanitization: No sanitization or validation of the
$ARGUMENTScontent is performed before it is passed to the tool chain, allowing potentially malicious instructions to reach the agent's processing logic.
Audit Metadata