create-mcp-skill
Warn
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends installing the 'mcptools' CLI using 'go install github.com/f/mcptools/cmd/mcptools@latest'. The GitHub user 'f' is not recognized as a trusted organization or well-known vendor in the security context provided.
- [REMOTE_CODE_EXECUTION]: The documentation suggests using 'bunx -y' and 'npx' to run MCP servers directly from the NPM registry (e.g., 'chrome-devtools-mcp@latest'). This pattern fetches and executes remote code at runtime, which can be vulnerable to package hijacking if versions are not pinned.
- [COMMAND_EXECUTION]: The skill's primary templates and examples utilize 'pkill -9 -f' to terminate processes matching a string pattern. This is an aggressive method of process management that can lead to unintended termination of unrelated system processes if the pattern is too broad or maliciously crafted via user input.
- [COMMAND_EXECUTION]: The 'Template Generator' section uses shell redirection ('cat > ... << EOF') to create executable script content on the local filesystem based on user-provided variables, which is a form of dynamic script generation.
Audit Metadata