autospec-review

Spec gap detection audits design specifications against linked open and closed GitHub issues using a formal taxonomy to identify missing requirements, inconsistent deliverables, and scope drift before implementation proceeds. It links each spec to related issues and classifies gaps systematically.

Regression issue filing writes a per-run CSV ledger of discovered gaps and creates high-priority [REGRESSION] issues when warranted. It runs a reviewer-validated remediation pass to reduce false positives and tightens acceptance criteria before routing work back through the issue queue.

Yes, spec auditing evaluates both open and closed GitHub issues against design specifications. It uncovers missing requirements, half-finished deliverables, and scope drift by comparing spec text against the issue tree and applying validation guardrails to minimize false positives.

Automated spec audits run after implementation batches using date filters and remediation flags. Running with a start date parameter audits all specs from that point onward, emits gap findings, and generates a regression issue queue routed back through the split workflow.

Yes, deterministic spec discovery and issue linkage are required. The audit applies taxonomy-based gap classification with validation guardrails by comparing design spec text against linked GitHub issue trees, so issues must be connected to the specifications beforehand.