Skills
Navigate
skills/redth/mauidevflow/maui-ai-debugging/Gen Agent Trust Hub

maui-ai-debugging

Warn

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: MEDIUMPRIVILEGE_ESCALATIONDATA_EXFILTRATIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill documentation includes instructions to use administrative privileges for installing system dependencies.
  • Evidence: The use of sudo apt install xdotool in references/linux.md for enabling key simulation on Linux environments.
  • [DATA_EXFILTRATION]: The skill provides commands that allow the agent to read sensitive application data and intercept network communications.
  • Evidence: Commands such as MAUI secure-storage get for reading encrypted app secrets, MAUI preferences list for accessing configuration data, and MAUI network for real-time HTTP traffic monitoring in SKILL.md.
  • [DYNAMIC_EXECUTION]: The skill enables the execution of arbitrary code and runtime modification of the application state.
  • Evidence: The cdp Runtime evaluate command for executing JavaScript in Blazor WebViews and MAUI set-property for live editing of native control properties in SKILL.md.
  • [COMMAND_EXECUTION]: The skill relies on a wide variety of system tools, build utilities, and automation frameworks to manage the app lifecycle.
  • Evidence: Extensive use of dotnet build/run, adb, xcrun simctl, open, osascript (for UI automation on macOS), and xdotool (for key simulation on Linux) across all skill files.
  • [EXTERNAL_DOWNLOADS]: The skill installs several global tools and NuGet packages, and includes a mechanism to update its own instructions from a remote source.
  • Evidence: Instructions for installing Redth.MauiDevFlow.CLI, androidsdk.tool, and appledev.tools, along with the maui-devflow update-skill command described in references/setup.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the target application's user interface, logs, and network traffic, creating a potential path for data-driven attacks on the agent.
  • Ingestion points: Visual tree data (MAUI tree), application logs (MAUI logs), and DOM snapshots (cdp snapshot).
  • Boundary markers: No explicit delimiters or warnings to ignore instructions within ingested data are documented.
  • Capability inventory: The skill possesses extensive capabilities for command execution (dotnet, adb), UI interaction (tap, fill), and dynamic execution (cdp eval).
  • Sanitization: No sanitization or validation of ingested content is described.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 24, 2026, 08:00 AM