Honor approved batch Apply scope and execution privacy

[Chris0Jeky]

Paper batch Apply still excluded shared-board proposals by other authors, higher-risk proposals and operations beyond five card creations, despite the maintainer D-4 ruling. Honor that ruling for live Approved proposals while keeping batch approval restricted, explicit Apply confirmation, the 500-item cap, revision/scope pins, per-item server authorization and receipts. Confirmation now also binds to the current reviewer so an account switch cannot inherit consent.

Single execute now conceals unreadable board and foreign boardless proposals with the same 404 response as missing proposals. Readable boards without Write still return 403. GP-06 records the trust-tier contract.

Refs #1307. Implements recorded D-4(a), (b), (c); the maintainer's personal dogfooding week remains open. No auto-approval or automatic board write is introduced.

Validation at 25b6fbd:

• Red-first widening tests: 7 failed / 19 passed. Account-switch regression separately failed before the identity pin.
• Scoped frontend approve/execute: 50/50.
• Full frontend: 391 files, 6,220 passed / 3 skipped.
• Full backend solution: Domain 1,631; Application 4,210; API 2,889 / 4 skipped; CLI 243; Architecture 28 / 1 skipped; Integration 36. Zero failures.
• Chromium: 1/1; a high-risk six-operation proposal creates zero cards before confirmation, then six cards after one explicit Apply, with a receipt.
• Typecheck, build, scoped ESLint, docs governance, Golden Principles, Markdown links and diff checks passed.
• Two fresh-context Terra high reviews: frontend consent and backend authorization, both SHIP. Only non-blocking controller-comment drift, tracked on REVIVAL-11: Risk-tiered review prioritization + explicit batch-confirm + model-derived confidence #1307.

Test setup corrections: process-local retired Gemini variables were removed for the Mock E2E run; two fixture failures were corrected to use numeric API enums and the cards endpoint. Persistent environment settings were untouched. Owned E2E servers stopped after the run. Hosted CI remains pending.

Docs: GP-06 updated. STATUS/MASTERPLAN/TESTING_GUIDE are leased by the existing-PR closeout lane; shipped-state reconciliation follows merge without concurrent edits to those files.

Sandbox limitation: single execute uses the configured read policy, which permits reads across existing boards in development sandbox. It can return 403 where batch execute returns 404; strict write gates remain enforced. This MEDIUM parity residual is tracked in #1307 (issuecomment-5600299037). Normal membership-based unreadable/missing concealment is covered by the API tests.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-09T10:37:43.347291Z	c3c2220	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[Chris0Jeky]

Independent reviews at 25b6fbd versus base 7ac99a7: Terra high frontend-consent lens SHIP; separate Terra high authorization/privacy lens SHIP. No CRITICAL/HIGH findings. Both noted the same non-blocking controller comments at lines 438/545 still describe the former single-vs-batch 403 distinction; tracked on #1307, no runtime effect. Round 1 complete.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 25b6fbd3f8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[Chris0Jeky]

Open-PR closeout coordinator: I am reconciling this new ready candidate into the current commit-preserving integration PR #2815, alongside the earlier ten reviewed heads. Please hold this source head and do not merge main independently while final combined qualification is running. I retain canonical docs, final interaction review and merge ownership; your source review/evidence remains scoped and preserved.

[Chris0Jeky]

Closeout integration has materialized 25b6fbd in #2815. I am classifying the new sandbox-visibility P2 against the explicit DevelopmentSandbox access contract before final qualification. Please hold source edits while this bounded interaction review completes; any confirmed blocker must be coordinated into the single integration repair batch.

[Chris0Jeky]

Base integration checkpoint: merged main at 64508d3, preserving commits. New head c3c2220. The scoped PR patch-id is identical to the independently reviewed patch; no new feature logic was added beyond the already reviewed palette base. 64 batch/palette tests and typecheck passed on the combined head. Existing full-suite evidence remains as recorded; fresh required hosted CI is pending. No merge claim until that gate passes.

[Chris0Jeky]

Delivered through merged integration PR #2815 at22029c760653519622de2e9bcdd86bb668a50d41. Read-only git merge-tree against this branch produces exactly the current main tree3a1a2b7bdbc3e6f844a411abd52a1c5f694bd458, so merging this PR would introduce no file changes. Closing the superseded PR without another merge. Original feature commits, local proofs and review comments remain available; parent issue residuals and human acceptance are not closed.

[Chris0Jeky]

Current-head checks are green. The source change is already delivered in #2815; this remaining base-merge head produces exactly the same qualified main tree3a1a2b7bdbc3e6f844a411abd52a1c5f694bd458. Finalizing the PR preserves its later ancestry without source/test/configuration changes. Broader #1307 acceptance and the triaged sandbox consistency residual remain open.

[Chris0Jeky]

[Codex lane release v2]
Recorded D-4(a/b/c) implementation is now on main through #2815, merge22029c760653519622de2e9bcdd86bb668a50d41. #2818 was closed as superseded after git merge-tree proved it adds no file changes to main (identical tree3a1a2b7bdbc3e6f844a411abd52a1c5f694bd458). Original scoped commits, full backend/frontend and Chromium proof, and independent reviews remain recorded on #2818.

D-4(d), the maintainer dogfooding week, remains unverified and is the blocking acceptance item. Sandbox response parity and stale XML comments remain explicitly tracked non-blocking residuals. No human box changed.

Owned worktree removed without force after clean/ignored inventory. Local logs, synthetic SQLite database/sidecars and Playwright test-results were preserved at C:\Users\jekyt\source\Taskdeck.claude\local\milestone-20260909\preserved-2818. Rebuildable caches only were discarded; original checkout and unrelated worktrees preserved.