adobe-edit-quick-cut
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided video files and form responses, creating a surface for indirect instructions to be ingested by the agent context.
- Ingestion points: User-uploaded video files and user-defined answers in the Step 4 Q&A form.
- Boundary markers: No explicit delimiters or specific instructions to ignore embedded content are provided for the ingested data.
- Capability inventory: The skill has access to Adobe video processing tools (
video_create_quick_cut,video_resize) and file upload capabilities. - Sanitization: There is no evidence of sanitization or validation performed on the user-provided content before interpolation into tool parameters.
- [EXTERNAL_DOWNLOADS]: Fetches processed video assets from Adobe's Creative Cloud infrastructure using temporary presigned URLs generated during the workflow.
- [COMMAND_EXECUTION]: In non-UI environments such as Codex, the instructions guide the agent to execute shell commands using the
curlutility to download output files to the local workspace.
Audit Metadata