dotnet-exception-handling

Pass

Audited by Gen Agent Trust Hub on Nov 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code during its investigation phase, which represents a standard attack surface for indirect prompt injection where malicious instructions could be embedded in code comments.
  • Ingestion points: The agent reads .NET source files (.cs) and project configurations (.csproj, *.sln) using ripgrep in Step 3 of the Investigation phase.
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the analysis of file content.
  • Capability inventory: The skill performs file system reads/writes, executes shell commands (rg, grep), and can orchestrate GitHub actions (issue and PR creation) if the fix-all flag is used.
  • Sanitization: Source code content is analyzed directly for pattern matching without intermediate sanitization.
  • [EXTERNAL_DOWNLOADS]: The skill references documentation and community articles from trusted and well-known sources including Microsoft's official documentation and abp.io. These are used for architectural guidance and do not involve the execution of remote scripts.
  • [COMMAND_EXECUTION]: The skill utilizes standard system tools like ripgrep (rg) and grep for project discovery and pattern matching. These operations are restricted to the local file system and are used for their intended purpose of static code analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 2, 2026, 04:52 PM