windbg-diagnostic-method
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- Command Execution: The skill utilizes a local PowerShell script (
validate-diagnosis-output.ps1) to verify the structural integrity of diagnostic reports. This script performs static analysis using regular expressions and JSON parsing to ensure reports meet defined formatting standards. - Indirect Prompt Injection: The validation script processes diagnostic output provided by the agent. While the script reads external content, its operations are limited to verifying document structure (such as headings and Mermaid diagrams) and parsing specific JSON metadata, which constitutes a low-risk processing boundary.
Audit Metadata