skills/im5tu/claude/flutter-duit-bdui/Gen Agent Trust Hub

flutter-duit-bdui

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the configuration of transport layers (HTTP and WebSocket) to fetch UI layouts, components, and logic from external remote URLs.
  • [REMOTE_CODE_EXECUTION]: The framework includes a 'ScriptingCapabilityDelegate' and 'scriptingManager' designed for 'Embedded script execution', allowing logic provided by a remote backend to be executed within the client application.
  • [COMMAND_EXECUTION]: The skill provides instructions to execute system commands such as 'flutter pub add' and 'flutter pub get' for dependency management during setup.
  • [PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection (Category 8) because it is designed to ingest and process untrusted data from remote servers.
  • Ingestion points: Remote JSON data is ingested from network endpoints via 'XDriver.remote' and various transport managers.
  • Boundary markers: No specific delimiters or validation instructions are provided in the documentation to isolate or verify server-provided JSON content.
  • Capability inventory: The framework allows for network requests, native platform interaction via 'NativeModuleCapabilityDelegate', and client-side script execution.
  • Sanitization: The documentation does not specify sanitization or verification procedures for the incoming server-driven layouts and embedded scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 11:37 PM