caveman-rtk-integration
Pass
Audited by Gen Agent Trust Hub on Apr 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the global installation of the
cavemannpm package (npm install -g caveman). While this is a standard development task, it involves downloading third-party code from the public npm registry.\n- [COMMAND_EXECUTION]: The instructions involve creating a custom Node.js wrapper script and granting it executable permissions usingchmod +x. This script is subsequently used to process templates and data.\n- [PROMPT_INJECTION]: The skill processes external data and interpolates it into templates to generate agent output, creating a surface for indirect prompt injection.\n - Ingestion points: Data is read from a JSON file provided via command-line arguments in the wrapper script.\n
- Boundary markers: No explicit delimiters or instructions are used to separate untrusted data from the template structure in the provided examples.\n
- Capability inventory: The generated output is intended to be consumed by an agent via RTK, and the wrapper script utilizes file system access and dynamic rendering logic.\n
- Sanitization: No input validation or sanitization is performed on the data fields before rendering.
Audit Metadata