prototype-pollution-advanced
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides specific payloads for achieving remote code execution via prototype pollution in common Node.js frameworks like EJS, Pug, and Handlebars.
- [COMMAND_EXECUTION]: The skill documents how to inject shell commands into child_process.spawn and child_process.fork through prototype pollution.
- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: Processes JSON bodies and URL query parameters in example exploitation flows.
- Capability inventory: Includes access to child_process modules and server-side template engine internals.
- Sanitization: Discusses techniques for bypassing proto filters using constructor.prototype and other variations.
- Boundary markers: Not applicable as it is a reference playbook for security practitioners.
Audit Metadata