find-untested-sources

Pass

Audited by Gen Agent Trust Hub on Jul 25, 2026

Risk Level: SAFE
Full Analysis
  • Static Analysis Heuristics: The skill identifies untested code by parsing source files for type declarations and matching them against identifiers found in test files. This process is entirely local and static, avoiding the risks associated with runtime code execution or complex build-time dependencies.
  • Secure Dependency Management: The skill utilizes well-known, versioned packages for code parsing. The C# engine uses Microsoft's official Roslyn libraries, and the polyglot engine uses standard tree-sitter language packs.
  • Data Privacy and Scoping: Analysis is restricted to the specific repository root provided by the user. The tool explicitly prunes sensitive or irrelevant directories—such as .git, node_modules, bin, and obj—to prevent accidental exposure of metadata, credentials, or build artifacts.
  • No-Exfiltration Design: The scripts output analysis results as structured JSON to the standard output and do not perform any network operations or external data transfers.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 25, 2026, 03:09 PM
Security Audit — agent-trust-hub — find-untested-sources