The Agent Skills Directory

[SAFE]: The skill is purely instructional, focusing on Dart language features. It promotes best practices for data validation and schema handling.
[INDIRECT_PROMPT_INJECTION]: The skill provides examples for processing untrusted data sources like JSON payloads and URI path segments.
Ingestion points: resolveAllowedAssetSubpath in SKILL.md and processUserPayload in examples/json_patterns.dart process external lists and maps.
Boundary markers: The instructions emphasize explicit validation using patterns and exhaustiveness checking.
Capability inventory: The provided code only performs data extraction and console logging; it lacks dangerous capabilities such as file system writes, network requests, or dynamic code evaluation.
Sanitization: The resolveAllowedAssetSubpath example explicitly implements a check against path traversal by validating that path segments do not contain '..' sequences.

dart-use-pattern-matching