v2.0.0 — Explicit setup and honest product claims

Retire the weighted recommendation/create/re-pick workflow and direct model changes to explicit source-backed CI setup. /legacy-setup now explains retirement; existing project history, Jenkins metadata and CI tokens remain accessible. Landing/sign-in, illustrative flow, privacy, chat labels and README now describe evidence and estimated accounting without cheapest/savings/quality claims.

Version: 1.5.0 → 2.0.0 (major) because a supported workflow is removed. Coordinate with backend 2.0.0; older weighted writes receive HTTP 410. B8–B12 live verification, reviewed aliases/rates and production rollout remain gated; no deployed pins or schedules change.

Validation: 157 unit + 21 integration tests; 54 desktop/mobile mocked browser checks (catalog 12, execution 22, usage 6, public/retirement 14), plus one required real local-backend smoke against disposable PostgreSQL. Typecheck/build pass. Lint: zero errors, 14 existing warnings. Desktop/mobile visuals inspected. Existing bundle-size and Node/tooling warnings remain. Self-review complete; history is preserved.

Published image: ghcr.io/steve-droid/driftplain-frontend@sha256:5c1f176326a1443ad726aef8a6116a7de1402bd3416dcda1872dbc8422b9b0ce. Release workflow 35987574510 succeeded. Anonymous Linux amd64 pull passed; OCI version 2.0.0 and revision 51e9adee05844040b34b8e6f9f01979b332b9c62 match reviewed main. Publication only; no deployment.

v1.5.0 — Selected-run usage and task results

Dashboard now shows the selected model actually executed, complete estimates separately from partial known charges, and unavailable/legacy coverage. Full-period totals remain independent of paginated run details and feedback. Reports, patch identities, named-test evidence and diagnosis/original failures render as their own task results; zero, unknown usage and rate provenance remain inspectable.

Replaces hypothetical-savings dashboard headlines while preserving legacy stored values with explicit limitations. Sample projects stay illustrative. Depends on the additive backend /projects/{id}/usage/v1 contract; deploy the compatible backend/migration first. No deployment or fixture seeding is included.

Validation: 157 unit, 21 integration and 52 desktop/mobile browser checks (22 execution, 12 catalog, 12 public/legacy, 6 usage). Typecheck, production build and diff checks pass. Lint has 14 existing warnings and zero errors; the existing bundle-size warning remains. Desktop/mobile result views visually inspected.

Release: 1.4.0 -> 1.5.0 MINOR for the new compatible selected-run usage/task-results dashboard. Fresh tags, Releases and GHCR versions checked. No existing tag/image changes or deployment pins.

Publication verified September 24, 2026: successful release workflow; empty-config anonymous GHCR pulls and matching OCI version/revision labels.

• ghcr.io/steve-droid/driftplain-frontend@sha256:e45fa8b7333b5d2fc71edac88b8a66822c707807066bf619faa1bbc29351da6d

v1.4.0 — Other custom-task setup

Other can now be configured from Set Up CI with an explicit single-call or OpenCode mode, an exact eligible model/evidence choice, editable literal templates/prompts, bounded inputs/resources, permitted paths and maintainer-owned validation. The preview explains report-only versus disposable-edit capabilities and keeps missing validation unverified.

Mode changes invalidate the pick while retaining its label and author text. Explorer/reload/edit and legacy transitions preserve intent; old named drafts remain readable. Custom drafts use a UTF-8 byte bound, remove stale overflow copies and display an overflow warning. Pending integrations remain unselectable; no runtime activation or deployment.

Validation: 154 unit and 21 integration tests; TypeScript, lint (14 existing warnings, zero errors), production build; Playwright execution 22, catalog 12 and public/legacy 12 checks, including desktop/mobile, keyboard selection, literal preview, both modes, mode invalidation and legacy edits. Self-review fixed prompt label stability and bounded-draft edge cases. Existing bundle warning remains.

Release: frontend 1.3.1 -> 1.4.0 (MINOR), a backward-compatible custom setup workflow. Backend Other Jenkins stages are delivered separately; agent 1.7.0 remains unchanged. Publishing does not change deployment pins.

Publication verified: workflow 35975390241 succeeded. Anonymous GHCR pull matched version 1.4.0 and main revision 235c120; digest sha256:d8a0e8fd9e5731ee260dd8797d022432cbe48f6e4c40b86a6f73d3c79e0b3711.

v1.3.1 — CI setup availability wording

Replace an internal implementation-slice label in the unavailable Other option with user-facing availability copy. Selection behavior and the four named-task contracts are unchanged.

Validation: typecheck and 11 focused execution tests pass. This is a copy-only correction after the B13 full verification.

Release: frontend 1.3.0 → 1.3.1 PATCH, correcting shipped interface wording. Preserve the published 1.3.0 tag/image; no deployment.

Published image verified anonymously: ghcr.io/steve-droid/driftplain-frontend:1.3.1 at sha256:0562e83922a4f8efeda79c6c2984e188e62b94ad9f8806d783749b49f7d41b6f. OCI revision a4f5c352faa8dd62a3a90fba1d86dfc259917b98 and version match reviewed main. Release workflow 35973497364 succeeded; packaged assets contain the corrected wording.

v1.3.0 — Named CI setup and direct model search

New CI setup now starts with a named task and direct search over verified eligible models. Users explicitly choose a provider deployment and source observation; comparable benchmark recommendations stay separate from supported unranked choices and empty runnable sets. Explorer links carry exact model/evidence intent through sign-in, draft restoration and project editing.

Add Python/Node test-environment inputs, diagnosis stage/log inputs and the optional repair scope. Profile changes invalidate the previous pick visibly; saves and restores revalidate exact selections. Reuse Jenkins metadata with versioned commands and in-memory one-time tokens. Same-profile re-picks preserve configuration/preferences. Legacy history, edit workflows and /legacy-setup remain compatible; Other authoring belongs to B14.

Validation: 143 unit tests, 21 integration tests, 14 named-setup desktop/mobile checks, 12 catalog checks, and 12 legacy/public browser checks passed. Typecheck, lint (14 existing warnings; no errors) and production build passed. Visual inspection fixed mobile fieldset overflow and unnecessary navigation snapshots; focus behavior is preserved. Browser success paths use synthetic verified fixtures only. No pending integration is activated and no provider call or deployment occurred.

Release: frontend 1.2.0 → 1.3.0 MINOR for compatible named setup/search workflows. Requires the B13 backend additive filters/setup seam (1.12.0). Publishing does not update runtime pins.

Published image verified anonymously: ghcr.io/steve-droid/driftplain-frontend:1.3.0 at sha256:0de414edf0e485a8783694518ebf8fb336ee1c985cfc21508100add43cd587cc. OCI revision/version match reviewed main. Release workflow 35973040252 succeeded. The subsequent immutable 1.3.1 release only corrects availability wording.

v1.2.0 — Public benchmark explorer

Add anonymous benchmark browsing, canonical model profiles, source-label discovery, full-server-catalog typeahead, and URL-persisted comparison of 2–4 models or exact unresolved source rows. Public reads bypass JWT/project probes; existing authentication, CI setup and project flows remain compatible.

Evidence keeps source/version/protocol/evaluator/snapshot/metric/coverage scopes separate. Users choose among multiple observations; no automatic maximum, composite score, inferred identity or runtime-support badge. Missing values remain missing. Shared benchmark explanations support focus, Escape, hover and touch. The explorer consumes the additive backend 1.4.0 public metadata contract.

Validation: 132 unit/component tests, 21 HTTP integration tests, 12 new desktop/mobile browser checks, 12 existing public/auth browser checks and 6 Google-auth browser checks. Typecheck, lint (no errors), production build and whitespace checks pass. Local real-API smoke covered all 19 frozen families, 13 CodeReviewBench rows, unresolved comparison/reload and narrow screens. Self-review fixes preserve comparison URLs, correct popover focus, keep home single-screen, and wrap long frozen revisions.

Release: 1.1.0 → 1.2.0 (MINOR) for backward-compatible public explorer functionality. Publish to GHCR only; deployment and production imports remain separate.

v1.1.0 — stable pre-refactor checkpoint

Stable pre-refactor checkpoint

September 22, 2026. This release preserves the existing Driftplain project before the benchmark-explorer and CI-selection refactor. It does not implement that refactor.

Repository commit: 6d902996090a55faa251d497bd86d19424f3ea8c. The tag is an immutable source checkpoint; no application code or runtime configuration was changed for this release.

Coordinated component versions

Repository	Tag	Commit
driftplain-backend	v1.1.0	270b4d0a1b89fe1ab8a7a69103837e7a45529e2b
driftplain-frontend	v1.1.0	6d902996090a55faa251d497bd86d19424f3ea8c
driftplain-gitops	v1.1.0	8f38c042dff87df8058616c66ee7271bd335bcda
driftplain-infra	v1.1.0	af06662812014d843d0fda3daaf520bc252a8e8e

Deployment and recovery

The app continues running on the home Kubernetes cluster. Existing GitOps image digest pins remain unchanged. Backend/frontend tags trigger image publication only; deployment requires a separate GitOps change.

Use the coordinated source tags for an interview checkout. For a runtime rollback, use the tagged GitOps home profile and its pinned images, and verify database compatibility or restore through the infrastructure recovery runbook. A Git tag is not a database backup. Do not apply the retired AWS platform as a rollback shortcut.

Existing historical v1.0.0 tags remain unchanged. The temporary unprefixed 1.0.0 checkpoint created during this session was replaced by v1.1.0.