skills/nvidia-nemo/speech/fix-issue/Gen Agent Trust Hub

fix-issue

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill retrieves issue descriptions and metadata from the official NVIDIA-NeMo GitHub repository via the gh issue view command.
  • [COMMAND_EXECUTION]: The skill utilizes standard development tools, including pytest for executing local test suites and git/gh for branch management and pull request creation.
  • [PROMPT_INJECTION]: The skill processes untrusted external data from GitHub issues. While this represents a surface for indirect prompt injection, the risk is mitigated by the instruction to manually reproduce the failure with a new test case and verify the fix locally before any external actions are taken.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 07:46 AM
Security Audit — agent-trust-hub — fix-issue