flutter-design-anti-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external Flutter source code to identify design defects, creating a surface where malicious instructions could be embedded in code comments or string literals intended for the reviewing agent.
  • Ingestion points: scripts/detect.dart.py reads files from the target directory, and the lint_plugin package analyzes the project's Dart AST.
  • Boundary markers: No explicit boundary markers or instructions to ignore embedded content are present in the detection output.
  • Capability inventory: The analysis tools are restricted to file system read operations for the purpose of source code scanning.
  • Sanitization: Content is filtered through specific Regex patterns and AST node selectors, which naturally restricts the data passed to the agent to relevant code structures.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 04:34 AM
Security Audit — agent-trust-hub — flutter-design-anti-patterns