qinghu-model-pose-set
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPERSISTENCEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides procedures to download the Node.js runtime from official servers and install the merchant's CLI tool from the public NPM registry. These downloads originate from well-known technology services and the process includes a SHA256 integrity check to verify the downloaded archive before extraction.
- [PERSISTENCE]: The setup instructions suggest adding the custom Node.js binary path to the user's shell profile (~/.bashrc). This ensures the runtime remains available across multiple terminal sessions, which is standard behavior for non-standard path installations.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data in the form of local image files or URLs to generate pose variations. 1. Ingestion points: Image paths are passed as parameters to the generation command in SKILL.md. 2. Boundary markers: No specific boundary markers or instructions to ignore embedded content are used. 3. Capability inventory: Command execution via qhkit and file system modification through installation scripts. 4. Sanitization: No explicit sanitization or filtering of image metadata mentioned.
- [COMMAND_EXECUTION]: The skill executes the qhkit utility using structured JSON arguments. The command structure is predefined, minimizing the risk of shell injection from user-provided prompts.
Audit Metadata