muapi-seedance-2
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to execute local shell scripts, specifically scripts/generate-seedance.sh and ../../../../core/media/generate-video.sh, to perform its primary video generation and management functions.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to build a technical Director Brief which is then passed as an argument to shell commands. 1. Ingestion points: User-provided descriptions of scenes, subjects, and actions in SKILL.md. 2. Boundary markers: None identified in the prompt templates to distinguish between user intent and instruction overrides. 3. Capability inventory: The skill has the capability to execute bash scripts with multiple flags including --subject, --intent, and --aspect. 4. Sanitization: There are no explicit instructions for the agent to sanitize or escape shell metacharacters within the user-provided text before interpolation into the script arguments.
- [DYNAMIC_EXECUTION]: The skill uses a pattern of dynamic command construction where shell execution strings are generated based on the mapping of user intent to technical parameters.
Audit Metadata