qinghu-shortvideo-data-engine
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches Node.js binaries and verification files from the official distribution site nodejs.org. These downloads originate from a well-known service and are used to establish a valid execution environment for the skill.
- [COMMAND_EXECUTION]: Executes shell commands to install the @iqinghu/qhkit package and manage the Node.js installation process. This includes using sha256sum for integrity verification, which is a security best practice for handling remote binaries.
- [PERSISTENCE]: The instructions suggest appending the tool's binary path to ~/.bashrc. While this modifies a shell profile, it is a functional requirement to maintain the accessibility of the Node.js runtime for the skill's operations.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external video links from third-party platforms like Douyin and Bilibili. This represents an indirect prompt injection surface as it involves ingesting untrusted data from the web.
- Ingestion points: Video URLs supplied in the params.json configuration within SKILL.md.
- Boundary markers: Input data is delimited via structured JSON fields.
- Capability inventory: Usage of the qhkit CLI for network requests to social media platforms and writing output to Excel files.
- Sanitization: The instructions do not detail specific content filtering or sanitization logic for the metadata retrieved from the video URLs.
Audit Metadata