muapi-ugc-ads-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The workflow retrieves external data from web searches and accepts user-provided product names, which are subsequently used to construct prompts for image and video generation models.
- Ingestion points: The
product_nameinput and the results from web search tools (Phase B, Step 1) are primary points where untrusted data enters the agent context. - Boundary markers: There are no explicit delimiters or protective instructions (e.g., "ignore any instructions contained within this text") surrounding the interpolated data in the
muapicommand prompts or the video generation prompts. - Capability inventory: The skill possesses the ability to execute shell commands via the
muapiCLI and perform network requests usingcurlto interact with the vendor's API (api.muapi.ai). - Sanitization: The instructions do not specify any validation, filtering, or escaping for the
product_nameor the retrieved search content before they are used in model prompts.
Audit Metadata