Quality Evidence

What this page proves

BoringToolsKit is moving from trust claims to visible operating evidence: formula notes, assumptions, source links, regression tests, browser QA, and correction paths.

Current public checks

• Site contract tests run before production deployments.
• Formula scripts that are present and expose a calc() entry point are loaded in a sandbox with default inputs and executed. A crash, an empty result, or an unresolved template token fails the check, as does a NaN in the output — except where the page itself reports N/A or NaN-safe for that input. Scripts that are missing or expose no calc() function are reported as skipped, not as passing. Reproduce it yourself: node tests/js/golden-smoke-all.js.
• Latest formula-execution run: 698 of 871 tool rows produced structurally valid output, 173 were skipped, 0 failed. Skip reasons: 127 result-invalid-defaults, 20 no-formula-file, 19 no-calc-function, 4 known-limit-throw, 3 known-limit-empty-result. Two additional non-tool regression assertions run separately and are not counted as tools.
• "Skipped" means no qualifying result was observed for that input set — it is not a pass, and it does not mean validation was correct.
• These checks confirm each script runs and produces output. They never compare a number against an expected answer, so a separate reference-case suite covers accuracy.
• Important calculators include formulas, examples, assumptions, related tools, and review dates.
• Portable actions share page URLs by default, not hidden inputs.
• Ad and sponsor modules fail closed unless real credentials are configured.

Inspect the evidence

• Public calculation tests
• Calculation-test suite on GitHub (clone & verify yourself)
• Correction history
• Local scenario workspace
• Product roadmap

Known limits

Some legacy tools still need deeper interpretation, scenario comparison, and more tool-specific source evidence. We are prioritizing high-stakes financial, health, generator, and energy tools first.