flutter-mcp-toolkit-guide
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run the
flutter-mcp-toolkit doctor --jsonCLI command to verify the environment and troubleshoot connectivity issues. - [REMOTE_CODE_EXECUTION]: The toolkit includes an
evaluate_dart_expressiontool, allowing the agent to execute code at runtime within the context of the connected Flutter application for debugging. - [PROMPT_INJECTION]:
- Ingestion points: The agent reads potentially untrusted data from the external Flutter application, including logs (
get_recent_logs), error envelopes (get_app_errors), and UI semantic data (semantic_snapshot). - Boundary markers: The instructions do not define specific delimiters or guidelines to ignore instructions that might be embedded within the application's output.
- Capability inventory: The agent possesses interaction capabilities such as UI manipulation (
tap_widget,enter_text), application state control (hot_reload_flutter), and shell execution (doctor). - Sanitization: There is no evidence of sanitization or structural validation for data ingested from the target application.
Audit Metadata