flutter-apply-architecture-best-practices

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines patterns for ingesting untrusted data from external sources through DataSources and ItemApiModel in SKILL.md. It provides instructions for the agent to utilize network requests via Dio, local database operations via isar, and native platform calls via MethodChannel. The guide does not specify the use of boundary markers or input sanitization when processing this raw external data, creating a potential surface for instructions embedded in data to influence the application's logic.
  • [DYNAMIC_EXECUTION]: The documentation encourages the use of dart:ffi (Foreign Function Interface) to call native C/C++ code. This allows for the execution of code outside the safety constraints of the Dart VM, which represents a vector for low-level system interactions if the loaded libraries are not strictly controlled.
  • [COMMAND_EXECUTION]: The skill recommends using MethodChannel and EventChannel for native platform integration. These capabilities allow the Flutter application to invoke native system functions and execute commands on the host operating system, which could be exploited if triggered by unsanitized inputs from remote data sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 07:04 AM
Security Audit — agent-trust-hub — flutter-apply-architecture-best-practices