muapi-product-showcase-video
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill takes user-provided text for 'ingredients_description' and 'brand_colors' and interpolates them directly into prompts for the 'bytedance-seedream-v4.5-edit' and 'seedance-v1.5-pro-i2v-fast' models.
- Ingestion points:
ingredients_descriptionandbrand_colorsinputs inSKILL.md. - Boundary markers: None present; inputs are placed directly into the natural language prompt string.
- Capability inventory: Executes
muapiCLI tools andcurlnetwork requests to external API endpoints. - Sanitization: No escaping or validation of user-provided text is performed before prompt interpolation.
- [COMMAND_EXECUTION]: The skill uses the
muapiCLI to perform image editing and video generation tasks, and provides instructions for the agent to usecurlfor API calls if CLI aliases are missing. - [EXTERNAL_DOWNLOADS]: The skill communicates with the external endpoint
api.muapi.aito process generation requests and poll for results.
Audit Metadata