flutter-app-icon-engineering
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides standard instructional content for Flutter developers. It correctly identifies required asset formats and directory structures for Android, iOS, macOS, Web, Windows, and Linux. All external references point to official documentation from Google, Apple, and the Flutter project.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface where the agent processes external project files.\n
- Ingestion points: The skill reads
pubspec.yaml,AndroidManifest.xml,manifest.json, andContents.jsonto determine project configuration.\n - Boundary markers: Not explicitly defined in the instructions; the agent is directed to read files to verify existing state.\n
- Capability inventory: The skill includes capabilities to modify file assets and execute build-related commands like
dart runandflutter build.\n - Sanitization: No explicit sanitization of input file content is described, as the skill operates on the assumption of a standard Flutter project structure.\n
- Contextual Note: This behavior is intrinsic to the task of icon engineering and does not constitute a security risk in this context.
Audit Metadata