linkpix-vidu-q3-clone
Fail
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: HIGHPERSISTENCEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PERSISTENCE]: The skill provides instructions to modify the user's shell profile (
~/.bashrc) to persist changes to thePATHenvironment variable. This is a persistence mechanism that ensures the custom Node.js installation remains active across sessions. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, specifically Douyin, TikTok, and other video sharing platforms via the
resourceUrlparameter. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the metadata or content of the linked pages. - Ingestion points:
resourceUrlparameter inqhkit video-inspire generate(SKILL.md). - Boundary markers: The skill instructs the agent to "rewrite" the script and obtain user confirmation before final submission, which acts as a manual boundary.
- Capability inventory: Subprocess execution of
qhkit,python, andnpx; network requests viacurlandnpm. - Sanitization: Relies on the agent to manually rewrite and the user to verify output; no automated sanitization is specified.
- [DYNAMIC_EXECUTION]: The skill generates and executes shell commands that include Python and Node.js code snippets at runtime.
- Evidence: Uses
python -cto execute a script for image compression using thePillowlibrary. - Evidence: Uses
npx --yes sharp-clito perform image resizing. - [EXTERNAL_DOWNLOADS]: The skill performs several external downloads during its environment bootstrap phase.
- Evidence: Downloads Node.js binaries and checksum files from
nodejs.org. - Evidence: Installs the
@iqinghu/qhkitpackage from the NPM registry, with optional use of thenpmmirror.commirror. - Evidence: Fetches the
PillowPython library from thetsinghua.edu.cnmirror if missing.
Recommendations
- HIGH: Downloads and executes remote code from: https://nodejs.org/dist/v22.22.3/SHASUMS256.txt - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata