api-contract-testing

Pass

Audited by Gen Agent Trust Hub on Jul 1, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references several external Dart packages (e.g., api_contract, pact_dart, json_serializable) from standard public registries for its core functionality.
  • [COMMAND_EXECUTION]: Instructions include standard build and installation commands (e.g., dart run build_runner build, flutter pub run pact_dart:install) which are typical for the Flutter/Dart development environment.
  • [DATA_EXPOSURE]: Includes examples of making network requests to third-party APIs (e.g., api.map.com.tw). It correctly uses placeholders like YOUR_API_KEY for sensitive parameters, following security best practices.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external API endpoints, which creates a potential surface for indirect prompt injection if those sources are compromised. However, the data is parsed as structured JSON for validation, which is the intended use case for this skill.
  • Ingestion points: SKILL.md (via dio.get calls to external APIs in testing examples)
  • Boundary markers: Not explicitly defined in the provided code snippets.
  • Capability inventory: SKILL.md (utilizes network requests via dio and local file writes for snapshots via File.writeAsStringSync)
  • Sanitization: Uses standard jsonDecode and regex-based JSONP parsing to process external responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 1, 2026, 06:02 AM
Security Audit — agent-trust-hub — api-contract-testing