javascript-expert
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The reference documentation includes educational examples demonstrating the use of Node.js child_process.exec and spawn for executing shell commands and managing external scripts.\n- [EXTERNAL_DOWNLOADS]: The module documentation references external libraries such as React and Lodash via the well-known esm.sh CDN and mentions the @js-temporal/polyfill package as a standard dependency.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and implement code based on external project files and user requirements, which establishes a potential attack surface for indirect prompt injection.\n
- Ingestion points: Project files like package.json, module system configuration, and user requirements specified in the Core Workflow.\n
- Boundary markers: No specific delimiters or instructions are provided to the agent to ignore potentially malicious instructions embedded in processed code or configuration files.\n
- Capability inventory: The skill documents and encourages the use of powerful Node.js capabilities including file system modification (fs/promises), network requests (fetch), and process execution (child_process).\n
- Sanitization: The instructions do not define sanitization or validation procedures for content extracted from external project files.
Audit Metadata