csharp-refactoring

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • Indirect Prompt Injection Surface: As described in SKILL.md, the skill is designed to ingest and modify C# source code files from a user's repository. This capability constitutes a potential ingestion point for indirect prompt injection, where instructions embedded in the code could influence agent behavior. The skill addresses this by enforcing a mandatory classification gate and using a capability inventory centered on standard validation tools like dotnet build and dotnet test. While it lacks explicit input sanitization, the requirement for successful builds and tests serves as a functional validation of the refactoring process.
  • Workspace Boundary Enforcement: In SKILL.md, the skill requires resolving the repository root using git rev-parse --show-toplevel. It explicitly prohibits the use of filesystem-wide searches or accessing paths outside the identified workspace, which reduces the risk of unintended access to sensitive files or configurations on the host system.
  • Trusted External Reference: A reference in SKILL.md points to an LSP configuration file hosted on a well-known version control platform within an official organization's repository. This external resource provides tooling metadata necessary for semantic refactoring and is used in a manner consistent with the skill's primary development purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 10:13 PM
Security Audit — agent-trust-hub — csharp-refactoring