wpf-best-practices

Warn

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Provides implementation patterns for modifying the Windows Registry (HKCU\Software\Microsoft\Windows\CurrentVersion\Run) to achieve application persistence (auto-start functionality).
  • [COMMAND_EXECUTION]: Includes instructions for the dotnet CLI to build and publish single-file executables, which is standard for .NET application distribution.
  • [EXTERNAL_DOWNLOADS]: References well-known and trusted .NET development tools and libraries from the official Microsoft and NuGet ecosystems.
  • [PROMPT_INJECTION]: The skill describes an Inter-Process Communication (IPC) mechanism using Named Pipes, creating a potential vulnerability for indirect prompt injection.
  • Ingestion points: The StartIpcServer method in SKILL.md accepts external input via a NamedPipeServerStream.
  • Boundary markers: The provided example does not include boundary markers or instructions to the agent to ignore embedded commands in the received messages.
  • Capability inventory: The described application possesses capabilities for file system access, network communication, and registry modification.
  • Sanitization: The code snippet lacks explicit validation or sanitization of the input received through the pipe before it is processed by HandleIpcMessage.
  • [SAFE]: Promotes secure local data storage by demonstrating the use of the Windows Data Protection API (DPAPI) for encryption.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 27, 2026, 06:03 AM