install-skill
Fail
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses Node.js to execute local scripts (
search.mjsandinstall.mjs) which handle file system writes and network requests. - [EXTERNAL_DOWNLOADS]: The skill fetches search results from the third-party domain
skills.shand repository metadata from the GitHub API (api.github.com). It downloads raw file content fromraw.githubusercontent.comwithout restricting downloads to a trusted list of repositories. - [REMOTE_CODE_EXECUTION]: The skill's primary function is to download and install new skills (comprising scripts and markdown instructions) from arbitrary GitHub repositories. Since these files are intended to be loaded and executed or followed by the AI agent, this creates a significant path for arbitrary code execution from untrusted internet sources.
- [COMMAND_EXECUTION]: The
install.mjsscript performs file system operations (creating directories and writing files) using input derived from external search results (skillName). A lack of sanitization on these inputs could theoretically allow directory traversal attacks, writing files outside of the intended directory.
Recommendations
- AI detected serious security threats
Audit Metadata