Quality Gates Report
This report separates reproducibility failures from unresolved industrial-engineering warnings.
Summary
| hard_failure |
total_gates |
passes |
warnings |
failures |
| False |
9 |
5 |
4 |
0 |
Gates
| name |
status |
severity |
message |
| required_report:reports/data_audit.md |
pass |
error |
reports/data_audit.md exists. |
| required_report:reports/model_report.md |
pass |
error |
reports/model_report.md exists. |
| required_report:reports/model_comparison.md |
pass |
error |
reports/model_comparison.md exists. |
| required_report:reports/decision_policy_report.md |
pass |
error |
reports/decision_policy_report.md exists. |
| required_report:reports/future_model_report.md |
pass |
error |
reports/future_model_report.md exists. |
| label_distribution_review |
warn |
warning |
Class 0 is the majority under the current label assumption. Confirm 0 = Failure, 1 = Production before operational use. |
| sensor5_dominance |
warn |
warning |
Sensor5 importance is 0.9856. Confirm it is a pre-decision raw measurement, not leakage. |
| sensor5_ablation_impact |
warn |
warning |
Removing Sensor5 drops macro F1 by up to 0.3275. Model selection is blocked until this feature is validated. |
| future_workflow_boundary |
warn |
warning |
Temporal future-failure report is a synthetic workflow demo, not real operational evidence. |
Interpretation
fail + error: fix before sharing or relying on reports.warn: known project risk that must stay visible.pass: gate condition is currently satisfied.