# Plugin release closure Status: DONE — locally owned release-closure work landed; remaining go-live blockers are concurrent, external, or final-RC evidence. ## Purpose Close the remaining **plugin release/product-readiness work that is not already owned by the concurrent `plugin-production-completion` thread**. This lane does not reimplement feature WIP that another thread owns. It turns the late-stage plugin into a truthful, reproducible, supportable release candidate for the first paying customer. Canonical baseline at plan creation: `origin/master` @ `13d43eaa7be47ef567161ca0533cf6cadff3635f`. ## Ownership boundary ### Explicitly NOT owned here The concurrent `docs/plans/2026-08-19-plugin-production-completion.md` lane owns all of the following. Do not touch its production/test files unless a concrete merge conflict requires a minimal reconciliation after that work lands: 1. connected-site credential provisioning in the remote E2E fixture; 2. real Worker E2E and release-regression proof for sync translation and async submit/poll/finalize/cancel; 3. ACF/Yoast accordion and translation-behavior reconciliation from `wt/modal-acf-yoast-accordions`; 4. page-fidelity implementation/regression recovery from `wt/page-fidelity-delivery` and the obsolete orchestration carrier; 5. manual-translation overwrite protection rebuilt from the preserved design/tests; 6. language-root switching, Yoast robots/sitemap ownership, Site Content schema/provisioning cleanup; 7. broad Chromium/Firefox E2E and final feature-gate execution; 8. stale E2E test-remediation work already tracked by `2026-08-15-e2e-failure-remediation.md`. ### External dependency — track, do not implement Stripe production clearance is pending outside this repository and is expected in roughly 2–3 weeks. Stripe checkout code is WIP/blocked on that clearance. This lane MUST NOT invent a replacement payment path, restore PayPal, or treat Stripe clearance as engineering work to be solved here. ### Existing capability — protect, do not rebuild The real translation lifecycle is an existing capability and has regression coverage (`real-translate.spec.js`, `SyncTranslationContractStandaloneTest.php`, `TranslateJobsCompletionStandaloneTest.php`, finalizer regression tests, bulk/cancel journeys). Treat a future failure as a regression. Do not redesign or rewrite this flow absent a failing current-master regression owned by the production-completion lane. ## Deliveries ### RC-01 — Correct the go-live tracker Update `GOLIVE.md` so it reflects current owner truth: - Stripe/checkout is `WAITING_EXTERNAL`, not an engineering unknown to be rediscovered. - Real translation lifecycle is `WORKS / regression-protected`; final release rerun is still required but implementation is not missing. - Remove Stripe and translation implementation from the list of work this lane should execute. - Keep feature WIP owned by the production-completion lane marked as concurrent/WIP rather than duplicating it here. - Keep final release evidence distinct from feature implementation. Acceptance: - no PayPal/license-key architecture reintroduced; - tracker distinguishes external wait, existing capability, concurrent WIP, and unverified release evidence; - `git diff --check` clean. ### RC-02 — Make customer-facing product copy truthful Rewrite `readme.txt` and `README.md` from current runtime/product facts rather than historical marketing claims. Required corrections: 1. Public product identity is **International Press Zone**, matching the plugin header and current admin UI. 2. Keep the actual runtime requirements from code/Composer: WordPress `>=6.0`, PHP `>=8.0`; do not claim PHP 8.3 as a minimum unless code changes first. 3. Remove unverified quantitative claims such as `10x faster`, `98% query reduction`, fixed cache-hit percentages, fixed latency, and unsupported scale/load numbers. 4. Remove hard-coded plan prices, support SLAs, “14-day free trial”, phone/24x7 promises, and package entitlements that are not sourced from the live commercial catalog. 5. Remove WordPress.org “search/install” instructions while this is a privately distributed commercial plugin; document ZIP install and signed in-product update flow instead. 6. Remove or qualify feature promises that are not in the frozen first-customer supported scope. Do not advertise WPML migration, WooCommerce/multi-currency, Elementor, Rank Math, white-label, team workflow, translation memory/glossary, or broad enterprise claims merely because historical code exists. 7. Keep only capabilities actually present and intended for the current commercial release: multilingual language management, translated posts/pages/strings, frontend localized routing/switching, AI translation through Press.Zone, bulk/job workflow, current account/connect surface, and integrations only once their concurrent acceptance lane is green. 8. Repair changelog chronology. Do not keep the fake/out-of-order historical `1.0.0 - 2026-01-26` entry in a `0.9.x` pre-release series. 9. Keep legal/license wording consistent with `composer.json` and the plugin header; do not silently switch license model in this task. 10. Never expose backend credentials, internal endpoints beyond what customers need, local development URLs, or test credentials. Acceptance: - grep finds no stale `Multilingual Press Zone`, hard-coded `$99/$299/$999`, `10x faster`, or free-trial/SLA claims in the two customer-facing files; - stable tag/header/version remain consistent; - no unsupported install path is promised; - docs remain useful enough for a paying customer to install and start onboarding. ### RC-02B — Retire stale legal/commercial templates The `legal/` drafts contain obsolete product identity, payment-provider, customer license-key, fixed-price, trial/refund, support-SLA, and categorical compliance claims. Replace them with architecture-aligned internal drafts/checklists rather than guessing new legal terms. Requirements: - use International Press Zone identity; - remove retired payment/license-key facts and hard-coded pricing/SLA/trial values; - describe current Connect/site-credential and translation-content data flows accurately at a high level; - record Stripe as pending external clearance; - remove categorical GDPR/compliance/security assertions not backed by current production evidence; - mark every legal file as draft/internal until qualified legal/business review; - make legal/privacy approval an explicit go-live requirement. Acceptance: no stale product/payment/license/pricing/SLA claims remain in `legal/*.md`; no document presents itself as approved legal advice or verified compliance. ### RC-03 — Freeze a first-customer supported-scope matrix Create `docs/launch/FIRST-CUSTOMER-SCOPE.md` with three states: - `SUPPORTED` — part of the first-sale promise and therefore must have acceptance evidence before release; - `CONDITIONAL` — supported only when the named integration/customer-stack condition is present and its dedicated acceptance lane is green; - `DEFERRED` — may exist in code but is not promised for this first commercial release. At minimum classify: - install/activate/update; - language CRUD/default language; - post/page translation; - string translation; - sync AI translation; - async/bulk translation and cancel; - frontend localized URLs/language switcher; - account/connect status; - Stripe purchase/billing as `CONDITIONAL / WAITING_EXTERNAL` until clearance; - ACF and Yoast as conditional on the concurrent reconciliation lane; - manual-translation protection and page fidelity as required once concurrent work lands; - WPML migration; - Site Content/templates/navigation/synced patterns; - WooCommerce/multi-currency; - Elementor / Rank Math; - team workflow/assignment/analytics; - translation memory/glossary; - multisite/broad third-party compatibility. The matrix is a release contract, not a roadmap. Anything `DEFERRED` must not be sold as production-ready in RC-02. Acceptance: - every major claim removed/kept in RC-02 maps to a matrix row; - no historical plan status alone makes a feature `SUPPORTED`; - concurrent WIP remains visibly conditional until its lane lands and passes. ### RC-04 — Add a deterministic release-readiness preflight Reuse existing tools; do not create a second distribution builder or deployment system. Implement a small non-mutating `tools/release-preflight.mjs` plus focused tests. It must: 1. require a clean Git checkout; 2. read and compare plugin header version, `IPZ_VERSION`, `readme.txt` stable tag, and distribution manifest expectations using the existing version parser rather than duplicate regex logic where possible; 3. verify required release tools/files exist (`factory-gate.sh`, `build-distribution.mjs`, `verify-local.mjs`, distribution manifest, rollback/release docs); 4. verify customer-facing docs contain the canonical product name and reject known retired architecture phrases in the active release surface (`PayPal`, customer license-key onboarding, `Multilingual Press Zone`) while allowing historical plan archives outside the release surface; 5. report concurrent blockers without mutating source: Stripe clearance, production-completion feature lanes, and final browser/runtime proof; 6. emit a concise machine-readable JSON result with `pass`, `blocking`, `waiting_external`, and `concurrent` groups; 7. exit non-zero only for deterministic local release defects, not for an external Stripe wait or still-running named concurrent lanes. This is a preflight, not the final release gate. The final release still uses the existing factory gate, distribution builder/verification, ZIP runtime verifier, sanctioned browser E2E, and backend update/package path. Acceptance: - focused Node tests cover version mismatch, dirty tree, retired customer-facing wording, missing release tool, and clean preflight; - no version bump or artifact build occurs during preflight; - no network access occurs. ### RC-05 — Define the final artifact/install/upgrade/update/rollback evidence contract Create `docs/launch/RELEASE-EVIDENCE.md` as the exact checklist/receipt schema for the final RC after concurrent feature work lands. It must require and record: 1. exact source commit and matching backend deployment commit; 2. full factory-gate receipt; 3. exact built ZIP path, version, byte size, SHA-256, runtime content digest, and manifest verification; 4. clean ZIP install + activation via the existing `verify-local.mjs` path; 5. upgrade from the immediately previous supported release with preservation checks for languages, translations, settings, connection metadata, and manual customer edits; 6. signed update/package check, download, tamper rejection, installation, and installed-version/hash proof against the canonical backend; 7. rollback artifact identity and tested restore procedure; 8. required Chromium+Firefox first-customer journeys from the production-completion lane; 9. explicit `WAITING_EXTERNAL` section for Stripe clearance so an unapproved payment processor cannot be accidentally marked complete; 10. one final GO/NO-GO line with no ambiguous partial pass. Acceptance: - every release-proof item in `GOLIVE.md` maps to this evidence contract; - the contract references existing tooling instead of inventing duplicate commands; - no dev1/production deployment is performed by this documentation task. ### RC-06 — Reconcile `GOLIVE.md` with landed evidence After RC-02..RC-05 and after any concurrent production-completion commits that land during this lane: - rebase/merge current `origin/master` through normal repository tooling; - update only criteria whose evidence changed; - mark concurrent items `WORKS` only from their landed receipts/tests; - leave Stripe `WAITING_EXTERNAL` until actual clearance/proof arrives; - recalculate the scoreboard from criterion lines, not prose estimates; - remove the old percentage-complete estimate if it cannot be mechanically defended. Acceptance: - `GOLIVE.md` is current against final landed master at the time of this lane's close; - no item is checked from code inspection alone where runtime evidence is required. ## Execution order 1. Land this plan to `origin/master` before implementation. 2. RC-01 tracker correction. 3. RC-02 customer-facing product truth. 4. RC-03 supported-scope matrix. 5. RC-04 release preflight + tests. 6. RC-05 release evidence contract. 7. Reconcile current `origin/master`, then RC-06. 8. Run applicable deterministic gates for files changed by this lane and land through `~/.claude/bin/land`. ## Stop conditions Do not wait for Stripe clearance in this lane. Do not start or duplicate page-fidelity, manual-lock, ACF/Yoast, Worker E2E, Site Content, language-root, sitemap, or broad E2E implementation. If those concurrent changes land, consume their receipts only during RC-06. ## Current receipt — 2026-08-19 - RC-01 implemented: GOLIVE distinguishes Stripe `WAITING_EXTERNAL`, existing regression-protected translation, concurrent feature WIP, and final release evidence. - RC-02 implemented: customer-facing README/readme rewritten to current International Press Zone identity and runtime facts; stale fixed pricing, trial/SLA, unsupported quantitative performance, WordPress.org-search, and broad unverified integration claims removed. - RC-02B implemented: stale legal/commercial templates replaced with current-architecture internal drafts; fixed pricing/trials/SLAs, retired payment/license-key instructions, and categorical GDPR compliance claims removed pending qualified review. - RC-03 implemented: `docs/launch/FIRST-CUSTOMER-SCOPE.md` created with SUPPORTED / CONDITIONAL / DEFERRED / WAITING_EXTERNAL release contract. - RC-04 implemented: non-mutating `tools/release-preflight.mjs` plus focused tests. Tests pass 6/6 for clean state, version mismatch, dirty tree, retired customer wording, retired legal wording, and missing release tool. - RC-05 implemented: `docs/launch/RELEASE-EVIDENCE.md` defines exact final-RC factory/browser/artifact/install/upgrade/update/rollback/Stripe receipt requirements using existing tooling. - RC-06 complete at `origin/master` `7df9ea27c2de64ff92917125a02ce5e9ce11774a`: no newer concurrent production-completion evidence had landed at close. GOLIVE baseline and scoreboard are reconciled to current master; page fidelity, manual protection, ACF/Yoast, Worker E2E, and other owned concurrent work remain explicitly outside this lane. - Clean committed-tree `node tools/release-preflight.mjs` returned `pass: true`, version `0.9.67`, zero deterministic blockers, two Stripe `WAITING_EXTERNAL` criteria, and one ACTIVE concurrent production-completion lane.