frontend-page-discovery
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Comprehensive review of the 10 threat categories revealed no security vulnerabilities.
- [PROMPT_INJECTION]: The skill instructions explicitly forbid speculative conclusions, requiring all output to be linked to file paths and line numbers as evidence. This design effectively prevents the agent from following potential instructions hidden within analyzed source code.
- [DATA_EXFILTRATION]: The skill only interacts with local project files for reading code and writing markdown documentation. It does not perform network operations or access sensitive system directories.
- [REMOTE_CODE_EXECUTION]: No external dependencies are fetched, and no remote code patterns were found.
- [COMMAND_EXECUTION]: The skill does not invoke any shell commands or system-level processes.
Audit Metadata