Quirk Coverage

A repository document, rendered for the site. View source markdown.

Generated at: 2026-07-30T12:38:02.000Z UTC · source: committed helm-expt evidence for this rendered repository document.

The catalog should not imply that every Helm behavior is fully understood. The quirk coverage audit records which behaviors are tracked, which are only visible in source scans, and which are known blind spots.

The generated audit is the current source of truth:

FileUse it for
data/quirk-coverage/summary.mdHuman-readable coverage summary by quirk axis.
data/quirk-coverage/coverage.csvSpreadsheet-friendly coverage table.

Each row has a coverage tier:

TierMeaning
tracked-and-surfacedThe behavior appears in generated chart/user data and has recipe or receipt evidence.
tracked-by-lock-not-front-doorThe behavior is recorded in locks or artifacts, but is not yet promoted into front-door tables.
partly-trackedThe behavior is visible, but one or more lifecycle or per-chart proof pieces is still missing.
disclosed-not-completeThe behavior is disclosed and gated, but not fully modeled or live-proven.
source-scanned-not-surfacedThe behavior is detected in source scans, but not promoted to chart facts or outcome tables.
not-scannedThe behavior is a known blind spot in the current scanner or data model.

Each row also separates source inventory from modeled support:

ColumnMeaning
source_top100_countPublic top-100 source-scan rows where the quirk was detected.
modeled_or_supported_countCurrent chart facts, recipes, receipts, or maintained work queues that model the quirk. The basis varies by axis and is described in the row.
source_top500_countPublic top-500 source-scan rows where the quirk was detected.

Do not read modeled_or_supported_count as the number of affected public charts. Do not read source_top100_count as proof that the quirk is supported.

Use the audit when adding chart facts, pain reports, variant paths, hook lifecycle routes, and top-100/top-500 analysis. A chart can still be usable when one of its axes is incomplete, but the incomplete axis should stay visible.

The generated summary also carries the current high-value counts for NGINX-like extension slots. These distinguish explicit catalog control points from broader source-scan signals such as raw tpl usage.

For the chart-level extension-slot list, use Extension Slot Coverage.

Regenerate and check the audit:

npm run quirk-coverage
npm run quirk-coverage:verify