adobe-create-social-variations
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use shell commands (
bash) to retrieve file size and MIME types for assets stored in the user-data upload directory. This is used for staging files before they are processed by external tools. - [DATA_EXFILTRATION]: The skill facilitates the transfer of local images and videos to Adobe's Creative Cloud storage for cloud-based processing. This behavior is core to the skill's purpose and is gated by an initialization check.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external media files and allows user-provided descriptions to influence AI-driven generative expansion and cropping. This ingestion of untrusted content represents a potential indirect prompt injection surface. * Ingestion points: User-uploaded files in
/mnt/user-data/uploads/and files selected via the Adobe file picker. * Boundary markers: None explicitly defined for isolating embedded metadata or content instructions. * Capability inventory: Shell execution (bash), network operations (PUT), and generative AI tool calls (image_generative_expand). * Sanitization: No explicit sanitization of file metadata or content is described prior to processing.
Audit Metadata