shiny-appfunctions
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a framework for exposing application logic to AI agents, creating an attack surface where untrusted user input from assistants is processed by code handlers. \n
- Ingestion points: Arguments provided by system assistants or LLMs are bound to C# request records as defined in SKILL.md. \n
- Boundary markers: The framework includes AppFunctionGate and IAppFunctionDelegate interfaces to implement authorization and allow/deny logic before function execution. \n
- Capability inventory: Function handlers have access to the application's registered services (e.g., databases, order systems) through dependency injection. \n
- Sanitization: The library uses JSON schema validation for parameter binding, and the documentation encourages developers to perform domain validation and throw specific exceptions for invalid input.\n- [EXTERNAL_DOWNLOADS]: The documentation references the NuGet packages Shiny.AppFunctions and Shiny.AppFunctions.Extensions.AI. These are official library components provided by the vendor for integrating .NET apps with AI assistants.
Audit Metadata