This pull request introduces a new infrastructure for reusable validation error definitions and implements several new built-in assertion methods for the validation library. The main goal is to allow both built-in and custom validation rules to reuse error definitions with clear defaults, ergonomic APIs, and efficient caching, especially for parameterized rules. The changes also provide new assertion methods for common validation scenarios and update documentation to reflect these new capabilities.

Validation Error Definitions Infrastructure:

• Introduced a new error-definition layer that distinguishes between reusable message templates and immutable validation error definitions, enabling efficient caching and reuse of rule identity and defaults. This includes support for both parameterless and parameterized definitions, a public static catalog for fixed built-ins, and a thread-safe singleton cache for parameterized rules.

New Built-in Assertion Methods:

• Added new assertion methods to the Light.PortableResults.Validation.Assertions namespace:
  • IsNotNull checks for null values and supports short-circuiting by default.
  • IsGreaterThan and IsLessThan check for comparative boundaries and use the new error-definition infrastructure, with optional short-circuiting. [1] [2]
  • IsIn checks for inclusive range boundaries using cached parameterized definitions.

Documentation and Guidance Updates:

• Added a comprehensive design plan (ai-plans/0026-error-definitions.md) explaining the rationale, acceptance criteria, technical details, and usage examples for the new validation error definition model and built-in assertions.
• Updated AGENTS.md to clarify API design guidance and document new validation behaviors, such as how Check<string?> normalizes nulls and how target composition is handled. [1] [2]

These changes make it easier to implement, reuse, and customize validation rules, and they lay the groundwork for more advanced and maintainable validation flows in the library.