Changelog
import { LinkCard } from ‘@astrojs/starlight/components’;
[0.8.0] — 2026-06-13
Section titled “[0.8.0] — 2026-06-13”Breaking
Section titled “Breaking”--cost-estimaterenamed to--effort-estimate— the flag outputs hours, not dollars; “effort” is more accurate. Update any scripts or CI pipelines that use--cost-estimate. The--hourly-rateflag is unchanged.
- Homepage email-capture / Loops signup section removed.
- Homepage version now read from
package.jsonat build time — can no longer drift from the published npm version. robots.txtsitemap URL corrected to/sitemap-index.xml(was/sitemap.xml, 404).- Audit sample corrected to real fixture output: 13 unique flags across 19 call sites, readiness 53/100, estimate 20.8h–40h.
Infrastructure
Section titled “Infrastructure”- Vitest timeouts raised to 30 s to fix flaky Windows/Node 22 CI runs.
npm publish --provenancere-enabled in release workflow.scripts/metrics/collect.mjsadded for tracking npm downloads and GitHub stats.
[0.7.0] — 2026-06-07
Section titled “[0.7.0] — 2026-06-07”flaglint audit --effort-estimate— directional migration-effort estimate in audit output. Produces a low/high hour range from automatable and manual-review call counts. Assumptions and disclaimer included in all report formats.flaglint audit --hourly-rate <rate>— engineering cost projection added to the estimate (costLow/costHigh). Requires--effort-estimate.- Migration readiness score —
flaglint auditnow prints a 0–100 score and progress bar showing the fraction of safely automatable calls, with grade (ready,moderate,complex, ornot-applicable). - Migration Readiness concept page — explains the ratio formula, grade thresholds, and the 5 manual-review categories.
- Effort Estimation CLI reference — documents the algorithm, default assumptions, minimum-hours floor, and hourly-rate behavior.
[0.6.0] — 2026-06-02
Section titled “[0.6.0] — 2026-06-02”flaglint audit [dir]— new command that generates a local flag debt audit report. Classifies detected LaunchDarkly Node.js SDK usage by migration risk based on call type, static analyzability, and migration complexity. Supports--format json,--format markdown, and--format html. No LaunchDarkly API key or credentials required.openFeatureClientBindingsinScanConfig— binding configuration is now included inScanConfigfor local migration planning and integration code paths.
isFeatureEnabled, React hook (useFlags,useLDClient), and wrapper function call sites are now included inmigrationInventoryafter scanning. Previously these appeared in scan reports but were invisible tomigrate --dry-runandmigrate --apply.providerfield removed fromScanConfig. The field was accepted but never read by the scanner. It remains inFlagLintConfigfor forward compatibility (v0.7).
Changed
Section titled “Changed”- README config table updated:
staleThresholdcorrected tominFileCount. - README migration table:
withLDConsumer()(Component)row updated —withOpenFeature()does not exist in the OpenFeature SDK.
[0.5.4] — 2026-05-29
Section titled “[0.5.4] — 2026-05-29”- Aliased named imports from the LaunchDarkly SDK (e.g.
import { init as ldInit }) are now correctly detected by the scanner.
[0.5.0] — 2026-05-28
Section titled “[0.5.0] — 2026-05-28”Changed
Section titled “Changed”- Narrowed scope to LaunchDarkly Node.js server SDK → OpenFeature migration. React/browser SDKs explicitly documented as outside current coverage.
migrate --applyguarded rewrite contract hardened: dirty git tree check, proven OpenFeature binding requirement, idempotency via range-content guard.
[0.4.0] — 2026-05-22
Section titled “[0.4.0] — 2026-05-22”- SARIF output via
validate --format sariffor GitHub Code Scanning integration. - Scan metadata (duration, file count) in all report formats.
[0.3.0] — 2026-05-18
Section titled “[0.3.0] — 2026-05-18”StalenessEvaluatorinjectable interface — enables external evaluators without touching scanner logic.
[0.2.0] — 2026-05-14
Section titled “[0.2.0] — 2026-05-14”Changed
Section titled “Changed”staleThresholdrenamed tominFileCount(breaking change).
View the full CHANGELOG.md on GitHub →