qinghu-product-swap
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes instructions to download Node.js binaries and associated checksums directly from nodejs.org, which is a well-known and official source for development tools.
- [COMMAND_EXECUTION]: The skill uses shell commands to extract downloaded archives via tar and install the @iqinghu/qhkit utility via npm. These commands are part of a routine installation and configuration process.
- [PERSISTENCE]: The skill suggests appending the local Node.js bin directory to the user's shell path in ~/.bashrc. This is a legitimate development practice to ensure that installed tools remain available across terminal sessions.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided prompts and image paths that are incorporated into tool parameters. 1. Ingestion points: User-provided prompts and file paths are used as arguments for the 'qhkit image generate' command. 2. Boundary markers: The skill instructions advocate for wrapping these parameters in structured JSON, which helps delineate untrusted user input from the command logic. 3. Capability inventory: The skill can execute file operations (uploading images) and network requests to the Qinghu AI API. 4. Sanitization: While explicit input sanitization is not detailed, the use of JSON structures for command arguments is a best-practice defensive measure.
Audit Metadata