skills/agentrhq/webcmd/webcmd-usage/Gen Agent Trust Hub

webcmd-usage

Warn

Audited by Gen Agent Trust Hub on Jul 4, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the core webcmd tool from official vendor sources, including the @agentrhq/webcmd npm package and the github.com/agentrhq/webcmd repository.
  • [REMOTE_CODE_EXECUTION]: The webcmd plugin install command documented in the skill allows for the installation of extensions from arbitrary GitHub repositories (e.g., github:user/repo), which facilitates the execution of unverified third-party code.
  • [COMMAND_EXECUTION]: The skill describes functionality to register and install arbitrary external CLI tools using webcmd external install and webcmd external register, which can introduce unmanaged binaries into the environment.
  • [COMMAND_EXECUTION]: The skill requests Bash(webcmd:*) permissions, allowing the agent to perform a wide range of operations including system configuration, plugin management, and browser automation.
  • [PROMPT_INJECTION]: The tool processes data from external websites and Electron applications, creating a surface for indirect prompt injection where malicious instructions embedded in web content could attempt to influence agent behavior.
  • Ingestion points: webcmd browser extract, webcmd <site> <command>, and adapter outputs.
  • Boundary markers: None mentioned in the provided documentation.
  • Capability inventory: Access to the webcmd CLI tool via bash.
  • Sanitization: No explicit sanitization or filtering of external content is described.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 4, 2026, 08:53 AM
Security Audit — agent-trust-hub — webcmd-usage