skills/dotnet/roslyn/formatting-log/Gen Agent Trust Hub

formatting-log

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • Command Execution: The skill automates the development workflow by executing shell commands, including dotnet test and a local PowerShell helper script. This is standard for its intended purpose of repository maintenance, but it assumes the execution environment and provided inputs are trustworthy.
  • Dynamic Code Generation: The Import-FormattingLog.ps1 script dynamically modifies the repository's source code by appending new test methods to FormattingLogTest.cs. While the script includes sanitization logic—such as validating that the test name is a valid C# identifier—the practice of runtime source code modification represents a security consideration that should be monitored.
  • Indirect Prompt Injection Surface: The skill is designed to ingest and process data from external sources, specifically zip archives and GitHub issue content. This introduces a surface where crafted filenames or issue metadata could potentially influence the agent's behavior or the content of the generated test code, although the script's validation logic mitigates the risk of direct command or code injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 10:37 AM
Security Audit — agent-trust-hub — formatting-log