Quality Studio connects code, findings and recorded review evidence in one workspace. Run reviews by scope and kind, inspect the reasoning, and version the results with your repository.
building in the open · library, CLI, API, and browser ship from one repository · nothing published to NuGet yet · Apache 2.0
Everything runs against a plain checkout. The analysis core is a packable library (nothing is published yet); the quality CLI drives scans, reviews, diffs, and reports; the API serves the tree, the overlays, the runs, and the usage; reports are exported as Markdown, HTML, JSON, or SARIF 2.1.0 with CI gates.
CI-ready exit codes
quality scan fails on stale reviews, quality diff --fail-on-regression on deterministic regressions, quality report --fail-under / --fail-on on score or severity — and the report is still published before the gate fails.
Scorecard
Grades per kind and level, finding counts by severity and lifecycle, fresh/stale/drift/missing distribution, file coverage, sensor posture. Missing coverage is visible, not a passing grade.
The project dashboard during today's sweeps: 81 open findings by severity, 13.3% review coverage, zero stale statements — and a security posture that reads scanner unavailable instead of a clean pass.
Handover
Handover to Agent Studio.
Quality Studio stays the engineer room. When a finding needs action, “make this a task” hands a lossless snapshot to Agent Studio through its normal task mutation path — one direction, no embedded quality dashboard in the cockpit, and a backlink for the way back.
Founded 2026-07-11. Core library, CLI, review API, and the Angular browser ship from one repository and are covered by CI. Open research is labeled as research: whether a code graph joins as a graphical meta layer is a time-boxed spike, not a decision. This page describes what exists.
