modellix-design
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill automatically manages the installation and update of the
modellix-clitool from the official npm registry (registry.npmjs.org). This behavior is documented and ensures the agent uses the current version of the vendor's orchestration software. - [COMMAND_EXECUTION]: The skill uses Python scripts to orchestrate shell commands, primarily
npmfor package management andmodellix-clifor executing AI generation tasks. These operations are implemented usingsubprocess.runwith argument lists, which prevents shell injection. - [INDIRECT_PROMPT_INJECTION]: The skill ingest user-supplied prompts for processing by AI models, which is its primary function.
- Ingestion points: User prompts are passed as data in JSON payloads within
SKILL.mdandscripts/invoke_and_poll.py. - Boundary markers: The skill utilizes structured JSON formatting to separate user-provided data from command parameters.
- Capability inventory: The skill has the capability to execute the vendor CLI tool, perform network requests to the vendor API, and download generated media files.
- Sanitization: Inputs are serialized into JSON structures before being passed to the CLI or REST API.
Audit Metadata