flutter-mcp-toolkit-custom-tools
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill documents the use of
fmt_evaluate_dart_expression, which enables the execution of arbitrary Dart expressions within the application context. This is presented as a method for simple data retrieval. - [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for building custom handlers that process external input (AgentArguments). This creates an attack surface where the agent's instructions could affect application state.
- Ingestion points: Tool and resource handlers registered via
AgentCallEntryinSKILL.md. - Boundary markers: The documentation recommends implementing strict JSON schemas and defining required properties to constrain input.
- Capability inventory: Registered handlers can interact with internal application logic, such as
CartRepositoryor feature flags. - Sanitization: The skill includes explicit safety guidance, advising developers to keep handlers thin and to avoid exposing secrets or performing unchecked I/O.
Audit Metadata