qinghu-video-watermark-remove

Fail

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: HIGHPERSISTENCEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [PERSISTENCE]: The skill instructions direct the agent to modify the user's shell startup profile (~/.bashrc) to persist changes to the system PATH. This allows the downloaded Node.js environment to remain active across different sessions.
  • [EXTERNAL_DOWNLOADS]: The skill downloads Node.js binaries from the official Node.js website and the @iqinghu/qhkit CLI tool from the NPM registry. While these are trusted or vendor-owned sources, the skill automates the download and installation of these external resources at runtime.
  • [COMMAND_EXECUTION]: The skill uses various shell commands, including curl, tar, sha256sum, and npm, to set up the environment and execute video processing tasks. It also instructs the agent to prompt the user for privilege escalation (e.g., sudo) if installation fails.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of video URLs and local file paths which are passed to the CLI tool.
  • Ingestion points: The urls parameter in the qhkit video-edit generate command found in SKILL.md.
  • Boundary markers: None. The skill lacks explicit markers or instructions to treat user-provided URLs as potentially malicious data.
  • Capability inventory: The skill can execute shell commands (qhkit, curl), perform network operations, and modify local configuration files.
  • Sanitization: No evidence of input sanitization or escaping was found for the user-supplied data before it is processed by the shell.
  • [DYNAMIC_EXECUTION]: The skill performs dynamic installation and execution of Node.js binaries and NPM packages during its setup phase. This includes uncompressing archives and executing the resulting binaries and scripts to prepare the environment.
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
Risk Level
HIGH
Analyzed
Sep 21, 2026, 04:10 PM