Outputs Reference

Config runs write a small artefact set intended for review and downstream reporting.

File	Purpose
`config.resolved.yaml`	Config with defaults made explicit.
`model_input.csv`	Effects and standard errors used for modelling.
`effect_preparation.csv`	Row-level source values, uncertainty provenance, transformations, exclusions, and warnings.
`readiness.csv`	Model-readiness status and messages.
`tracker_issues.csv`	Tracker quality warnings/errors.
`effect_summary.csv`	Pooled mean posterior summary.
`future_true_effect_summary.csv`	Latent true effect for a comparable future study.
`analysis_report.md`	Analyst-facing summary of reportability, diagnostics, readiness, and headline effects.
`diagnostics.csv`	Sampler diagnostic checks.
`prior_diagnostics.csv`	Approximate prior-information warning for `mu` plus `tau` prior-predictive checks.
`ppc.csv`	In-sample study-level posterior predictive checks.
`run_status.json`	Machine-readable run outcome, including `reportable`.
`outputs_manifest.json`	Machine-readable artefact list.

Optional files:

File	Written when
`uncertainty_sensitivity.csv`	`sensitivity.uncertainty: true`
`prior_sensitivity.csv`	`sensitivity.prior: true`

Prefer adding columns over renaming or removing columns. If a public artefact changes, update fixture tests and docs in the same change.