agentclash-challenge-pack-input-sets

YAML input sets for AI agent evaluation require a strict schema defining case keys, payloads, inputs, and expectations. This structure enforces validation rules to ensure test cases are stable, observable, and compatible with the evaluation runner.

Test cases for CI gates and regression suites are categorized by run-intent, organizing them into smoke, CI, full benchmark, regression, or edge-case suites. This categorization optimizes evaluation speed and coverage across different testing scopes.

Deterministic input sets for regression testing are created by standardizing YAML payloads and enforcing strict schema validation. This ensures unique case keys, valid references, and observable expectations are established before cases are published.

YAML input validation fails when cases do not meet strict requirements like unique keys, valid references, or observable expectations. Enforcing these schema rules prevents configuration errors and ensures compatibility with the evaluation runner.

Yes, you can define challenges, inputs, and expectations using the required YAML structure without referencing the source repository. The exact schema provides structured case design to maintain evaluation consistency and observability.