auto-skill-guide
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the @matrixy/auto-skill package via npm and npx. These resources belong to the developer's package namespace and are required for the plugin's functionality.
- [COMMAND_EXECUTION]: The provided install-cli.sh script executes shell commands to detect the Node.js environment and automate the global installation of the CLI tool. It includes a user confirmation prompt unless an environment variable is set to bypass it.
- [INDIRECT_PROMPT_INJECTION]: The plugin observes tool call sequences to generate future instructions. This represents a potential surface for indirect injection if tool sequences are influenced by malicious data.
- Ingestion points: Tool usage history recorded via PostToolUse hooks.
- Boundary markers: The system uses specific delimiters (e.g., 'SKILL LOADED' and 'END OF SKILL') when injecting generated instructions into the session context.
- Capability inventory: Includes file system writes to ~/.claude/skills/auto/ and execution of generated patterns.
- Sanitization: Employs a manual review and approval workflow (/auto-skill:review) which requires the user to preview and approve patterns before they are promoted to active skills.
Audit Metadata