muapi-freeze-effect-video
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill interpolates user-supplied text directly into the generation prompt for the video model, which could be used to influence the output if malicious strings are provided.
- Ingestion points: User-provided inputs
scene,freeze_moment, andclosing_lineare used in the prompt. - Boundary markers: There are no explicit delimiters or instructions to the model to ignore potential commands within these inputs.
- Capability inventory: The skill utilizes the
muapiCLI andcurlfor external processing. - Sanitization: No input validation or escaping is implemented for the text inputs.
- [COMMAND_EXECUTION]: The skill relies on the execution of the
muapicommand-line interface to generate video content. - Evidence: Instructions for the executing agent to call
muapi video from-imageandmuapi predict wait. - [DATA_EXFILTRATION]: The skill makes network requests to an external API to facilitate video generation.
- Evidence: Fallback instructions to use
curlfor POST requests tohttps://api.muapi.ai/api/v1/<endpoint>.
Audit Metadata