flutter-design-system-from-design-md
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access behaviors were detected. The skill's operations are confined to local code generation and project auditing based on provided design documents.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted design specifications (e.g.,
DESIGN.md) to generate implementation code, which is a standard ingestion surface for this type of development tool. The associated risk is minimal as the agent's capabilities are focused on formatting, analysis, and testing within the local development environment. - Ingestion points: Repository files including
DESIGN.md,pubspec.yaml, assets, and existing source code. - Boundary markers: None explicitly defined to separate design intent from instructions.
- Capability inventory: File system read/write access and execution of local Flutter development tools (
flutter format,flutter analyze,flutter test). - Sanitization: Not explicitly implemented; the skill relies on detailed architectural and implementation constraints to govern output.
Audit Metadata