comfy-minimax-modal
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands using
modalanduvto manage serverless infrastructure and run inference tasks. These are standard operations for the described workflow.\n- [REMOTE_CODE_EXECUTION]: The skill usesmodal runto execute repository-resident scripts (e.g.,modal/minimax_h3.py) on remote GPU infrastructure provided by Modal.\n- [EXTERNAL_DOWNLOADS]: The skill initiates the installation of themodalCLI tool and the selective download of large machine learning model files required for video generation.\n- [INDIRECT_PROMPT_INJECTION]: The skill contains a surface for potential indirect prompt injection attacks.\n - Ingestion points: User-supplied text prompts are interpolated into shell commands via the
--promptargument as shown inSKILL.md.\n - Boundary markers: There are no explicit delimiters or boundary markers used to wrap or isolate the user-provided prompt from the surrounding command context.\n
- Capability inventory: The skill has the capability to execute shell commands, manage remote infrastructure, and write output files to the local workspace.\n
- Sanitization: The instructions do not document any sanitization, escaping, or validation logic for the input prompt strings.
Audit Metadata