research-evidence-harness

Validating research experiment bundles requires applying dry-run validation, schema checks, and objective or metric calculations to generate project-local evidence artifacts. This process ensures dry-run evidence separation and prevents premature live claims before deployment.

Creating a project-local run schema involves converting planned research protocols into standardized data structures. This standardization validates the protocol's claims and data organization before any experiment execution begins.

Validating JSONL fixtures for dry-run evidence requires an append-only validator that checks fixtures against the project-local run schema. This process produces validation reports to ensure data integrity and maintain separation from live experimental data.

Generating publication-readiness summaries requires validated dry-run fixtures and result summaries. This evidence gating produces readiness indicators that guide lifecycle transitions and support publication decisions by verifying that claims meet governance requirements.

Yes, projects without existing governance schemas can initialize the validation workflow by providing the research project root. The system then generates the initial run schema, JSONL validator, and evidence artifacts from the provided protocol definitions.

Automated schema validation for research protocols is not suitable when live experiment data and dry-run evidence cannot be strictly separated. Projects lacking defined claims, protocols, or experiment bundles will not benefit from dry-run validation and evidence gating workflows.