ryota-murakami · Oct 23, 2026 · Oct 22, 2026 · Oct 22, 2026 · Oct 22, 2026 · Oct 22, 2026
diff --git a/AGENTS.md b/AGENTS.md
@@ -53,15 +53,15 @@ The driving agent owns state, cycle selection and completion claims; GStack supp
 
 **Background work.** `node extension/scripts/watch-background.mjs` (from the directory Claude Code started in, else with `--project <that directory>`) lists the current Claude Code session's background tasks and subagents (its `CLAUDE_CODE_SESSION_ID`), each with the time of its last output, and marks those silent for more than 20 minutes; a finished or stopped subagent is never marked, a shell's row stays marked after it ended (its task notification is the authority). Stop a review subagent that has written nothing for 20 minutes and dispatch it again on a narrower scope; dispatch a subagent that the API's safety filter stopped again at once, with the same brief, on the fallback model. While waiting longer than a few minutes, tell the owner what is awaited and since when.
 
-**Ship review limit (owner, 2026-10-18; 2026-10-23).** The installed `/ship` runs at most three fixing cycles: no fourth cycle, no counter reset after a handoff, no round relabeled a repair. When the third cycle still leaves findings (`converged:false`), the owner's rule of 2026-10-23 applies, in a feature cycle and a bug-fix cycle alike: fix every finding of the last round, Minor and below included (so a Critical one is fixed before the merge in both, as the delivery rules above require); record only the Critical and Major defects still open in a GitHub Issue (the next cycle's tasks); then merge after required CI and the exact-HEAD gate, reporting `converged:false` in the PR. This overrides the no-carry rule for those recorded defects only.
+**Ship review limit (owner, 2026-10-18; 2026-10-23).** The installed `/ship` runs at most three fixing cycles: no fourth cycle, no counter reset after a handoff, no round relabeled a repair. When the third cycle still leaves findings (`converged:false`), the owner's rule of 2026-10-23 applies, in a feature cycle and a bug-fix cycle alike: fix every finding of the last round, Minor and below included (so a Critical one is fixed before the merge in both, as the delivery rules above require); record only the Critical and Major defects still open in a GitHub Issue (the next cycle's tasks); then merge after required CI and the exact-HEAD gate, reporting `converged:false` in the PR. This overrides the no-carry rule for those recorded defects only. The limit is `/ship`'s alone (owner, 2026-10-23): one `/ship` round is long, and running it to full convergence would stall delivery. The defects it records are those of `/ship`'s review rounds that the pull request cannot close, such as a Major finding whose fix or proof needs a root cause still unknown or a native check. A Critical one never merges open: a Critical finding the pull request cannot close holds the merge until it is fixed, as the cycle rule and the delivery rules require. The local CodeRabbit CLI and the hosted CodeRabbit review have no round limit: their findings follow the cycle rule at every review, as do failures of the gate, CI and native checks.
 
 **Skills.** Read each installed SKILL.md at invocation; this map guides and copies no workflow. Premise or final behavior: `/office-hours` or `/plan-ceo-review` (owner decision). Lasting API, MCP or integration change: `/plan-devex-review` (capability, error, lifecycle contract). Flow, layout, hierarchy, interaction: `/plan-design-review` (UX and state decisions). Architecture, concurrency, storage, meaningful code: `/plan-eng-review` after UX/API amendments (bounded implementation and test plan). Combined review on owner request: `/autoplan` (CEO → conditional Design → conditional DX → Eng). Unexplained failure: `/investigate` (root cause, new hypothesis). Working browser or webview flow: `/qa` or `/qa-only` (observed behavior, fixes). Implemented UI: `/design-review` (real visual and interaction evidence). Ready diff: `/review` (revision-bound findings, coverage limits). Releasable PR: `/ship` (actual commit, PR, checks). Merge or deploy: `/land-and-deploy` if available, else the established extension delivery. Changed public behavior: `/document-release`. Pause or handoff: `/context-save`, then `/context-restore`, reconciled with the repository and Issue. Ordinary coding is the agent's work; documentation-only changes need only the applicable reviews; a failed experiment may inform the next cycle before release. Browser QA, Reader DOM checks included, proves no native pixels, DAP targeting, MCP integration or installed VSIX behavior. Report a missing tool as missing coverage.
 
 **Skills under the cycle rule** (each runs as installed). `/spec` plans a feature increment from PRODUCT_GOAL.md and the gap audit, or writes a bug-fix Issue from the previous closing comment: one task and acceptance criterion per carried defect, nothing ranked or selected. Plan reviews may change how a carried defect is fixed, never remove, merge away or defer one. In review steps (`/review`, `/ship`) a feature cycle answers non-critical findings Skip and applies no automatic fixes of informational findings; a bug-fix cycle answers none Skip. In `/ship`, the coverage audit's tests are committed; the plan completion audit counts an open bug-fix task NOT DONE and stops (no ship anyway, defer or drop); the TODOS step writes no TODOS.md; documentation-audit blockers are fixed before the commit. A test gap, deferral, non-blocking documentation debt or skill follow-up is carried in a feature cycle and fixed before the merge in a bug-fix cycle. A feature cycle's closing comment names every carried defect, and the bug-fix Issue comes before other work; a bug-fix cycle's states that nothing is carried, and `/spec` starts a feature cycle; when the ship review limit recorded defects, the closing comment names them and the next cycle is a bug-fix cycle.
 
 **Invocation.** Pass the Issue, phase and bounded scope explicitly. PRODUCT_GOAL.md is the destination, not permission to build everything. Retired G08/G09 never return through an old prompt. Give design skills `## Design target` and the PNG. Honor scope gates without re-asking authorized routine choices or passing agent-chosen targets off as the user's; keep required reports; one approved recommendation authorizes no unrelated change. Shared-skill and Wiki-generation scope gets DX and engineering review; discovery, loaded identity, runtime connection and AI-owned results are separate checks; the Wiki works without runtime or debug MCP.
 
-**CodeRabbit and OCR.** [.coderabbit.yaml](.coderabbit.yaml) holds policy and exclusions: keep its guidelines current, validate with `coderabbit config validate .coderabbit.yaml`, and with `--config` pass it beside the extra context. Required: the local CodeRabbit CLI, findings triage, CI and `open-code-review-delegate` (OCR picks files and rules, the host reviews; each eligible file reviewed or skipped with a reason; default exclusions reported apart; no independent external review). Review against the implementation base; record completion, source cut, covered files, findings. Hosted CodeRabbit findings fall under the cycle rule; if it did not run, report "not run", never APPROVED. Missing or private provider context is a coverage limit. Each user-visible behavior that a CHANGELOG entry names has a sentence in the packaged README (`extension/README.md`), as `.coderabbit.yaml`'s `CHANGELOG.md` instruction checks; a fix that changes a user-visible behavior changes that README in the same commit, and fixers' briefs say so. Claude Code CLI review stays owner-skipped and is no native chat acceptance.
+**CodeRabbit and OCR.** [.coderabbit.yaml](.coderabbit.yaml) holds policy and exclusions: keep its guidelines current, validate with `coderabbit config validate .coderabbit.yaml`, and with `--config` pass it beside the extra context. Required: the local CodeRabbit CLI, findings triage, CI and `open-code-review-delegate` (OCR picks files and rules, the host reviews; each eligible file reviewed or skipped with a reason; default exclusions reported apart; no independent external review). Review against the implementation base; record completion, source cut, covered files, findings. The hosted CodeRabbit review is triggered on each pull request (owner, 2026-10-23): wait for it beside CI and fix its findings under the cycle rule at every review, with no round limit (the ship review limit is `/ship`'s alone); when a rate limit or a failed trigger keeps it from running, report it "not run", never APPROVED, and go on with the local review, which runs every cycle either way. Missing or private provider context is a coverage limit. Each user-visible behavior that a CHANGELOG entry names has a sentence in the packaged README (`extension/README.md`), as `.coderabbit.yaml`'s `CHANGELOG.md` instruction checks; a fix that changes a user-visible behavior changes that README in the same commit, and fixers' briefs say so. Claude Code CLI review stays owner-skipped and is no native chat acceptance.
 
 **Status:** ready → planning → implementing → validating → closed, or blocked with its condition and unblocking action (say why work moves back). **Result:** pending, demonstrated, falsified, inconclusive. **Delivery:** not-applicable, local-only, committed, PR-open, merged, packaged, published, installed-verified. The three are independent.
 

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,34 @@
 # Changelog
 
+## [0.2.37.0] - 2026-10-23 (development candidate)
+
+Cycle 0050's second pull request ([Issue #118](https://github.com/laststance/codebook/issues/118)): a test replay's Application pane and Study section. The recorded-source tab and the moves between moments come with the third.
+
+**The Application pane of a test replay.**
+
+- **Your place in the output is kept.** The choice to follow the latest output and the line you read are kept per replay by Codebook, not by the page, so a rebuilt pane, a Retry, a presentation change or a page VS Code reloads opens where you were, with **Jump to latest** when you were reading higher up. Scrolling up stops following and Jump to latest follows again, decided by the scroll position alone. While you read higher up, a later moment that drops the earliest kept lines leaves the line you read where it stood.
+- **Show recorded output brings a closed pane back above the source.** It opens in a group of its own above the recorded-source tab's group, and no other pane moves; when your editor kept the emptied group (`workbench.editor.closeEmptyGroups` off), the pane returns into it.
+- **Opening another test replay** leaves the replay you see as it is until the other one shows, and the loading line names the replay being opened. A view the page did not take is sent again.
+- **Output reads as the test printed it.** Color sequences and terminal strings that the recording split across two pieces are removed whole, an unterminated one removes only itself, and a recording full of them is cleaned in time linear in its length. Lines with emoji count each emoji once, and the muted lines read `1 earlier character not shown` and `1 Node inspector line hidden` in the singular.
+- **Faster in long recordings.** Moments in a burst are drawn once per frame, newest only, unchanged lines keep their place on the page, and the output is read once per opened replay.
+
+**The Study section of a test replay.**
+
+- **No hidden reordering.** Bidirectional and terminal controls are removed from every text the section, the Application header, Saved studies and the learning summary print: test titles and files, output and result lines, stop paths, frames and values, and the learning material's question, title, summary, participants and steps.
+- **Why a capture is partial or interrupted** shows under the status, on one line in test replay words.
+- **Inspect in native debugger beside Show Variables.** While the debugger shows the stop and its values are not loaded, the row says `Debugger stop selected`; once the debugger no longer shows the stop (you ended its session, or it could not start), **Inspect in native debugger** there starts it again. The button follows the selected stop, also while the event list shows another page.
+- **Quick moves add up.** Two quick Next presses move two stops, and a move that fails, or that a Close, a reopen, Return to Live or a browser recording cuts off, gives its place back, so the next move starts from the moment shown.
+- **Keyboard.** A long title is a tab stop that shows whole on focus, the Application pane's output region is a tab stop you can scroll by keyboard, a control that leaves during Play hands the focus to the transport, and a control you reach with Tab never hides under the sticky transport.
+- **The sequence diagram** wraps names and labels at word boundaries, also Japanese and Chinese text, and scrolls by keyboard to its last participant.
+- **Test replay words.** Saved studies rows read `Test replay · …`, and refusals say test replay where they said Node history: `Test replays cannot be deleted here yet.`, `There is no next stop in this test replay.`, `The replay debugger was stopped. Select this stop or use Inspect in native debugger to restart it.` (said by a seek or Play that reaches a stop after you stopped the debugger; Next stop, Previous stop or Inspect start it again).
+- **A browser recording opened while a test replay is selected** retires the test replay, and the Application pane shows the recording's own pages.
+
+**Another window of the same workspace.** A test history that another live window records is listed and opened as recording, not interrupted, and its journal is read only where it grew; once that window seals it, it reads with its sealed status, and if that window ends without sealing it, as in a crash, it reads interrupted. A writer record that could not be read for a moment is judged again a few seconds later.
+
+**Compatibility.** The MCP API stays 1.8. A history being recorded now carries a small writer record in its folder, removed when it seals; earlier versions ignore it, and test histories, learning materials and reader positions open in both directions unchanged.
+
+**Observed natively.** Show recorded output after closing the pane, in VS Code 1.142.0 and Cursor 3.26.16 on macOS: the pane returns alone above the source and nothing else moves. In VS Code, an output reached from a result read away from its end keeps following off and Jump to latest. Windows and Linux are not run.
+
 ## [0.2.36.0] - 2026-10-23 (development candidate)
 
 Cycle 0050 is a bug-fix cycle ([Issue #118](https://github.com/laststance/codebook/issues/118)): it fixes the defects Cycle 0049 left, in three pull requests merged in order. This version holds the first: the test run's start, stops and refusals, saving and storing test replays, causes that name no private path, the bundled skill and the tests that failed under load. The Application pane and the Study section come with the second, the recorded-source tab and the moves between moments with the third.
@@ -23,7 +52,7 @@ Cycle 0050 is a bug-fix cycle ([Issue #118](https://github.com/laststance/codebo
 - **A replay debugger that cannot start** (a declined workspace trust, the capture timeout) no longer fails the open: the replay opens on its moment, the Study pane says `The replay debugger could not start: <cause>. Select this stop or use Inspect in native debugger to restart it.`, and Retry starts it again. A debugger you stopped stays stopped when you move to an output or the result.
 - **`codebook_playback` refuses a test replay** with `PLAYBACK_PROFILE_UNSUPPORTED`, also one of an earlier session that this window has not listed yet.
 
-**Causes name no private path.** Every cause an agent or a learner reads (MCP answers and refusals, the run resource, `codebook_status`, Saved studies, the book catalog, notices, Output › Codebook) shows the extension's storage folder as `<storage>`, the opened repository as `<workspace>` and the home directory as `~`, also under their real paths. A long cause is cut between characters and ends with `… [cut]`; the run resource shows at most 300 characters of it.
+**Causes name no private path.** Every cause an agent or a learner reads (MCP answers and refusals, the run resource, `codebook_status`, Saved studies, the book catalog, notices, Output › Codebook) shows the extension's storage folder as `<storage>`, the opened repository as `<workspace>` and the home directory as `~`, also under their real paths. A long cause that a tooltip, a notification, the run resource or an MCP message shows is cut between characters and ends with `… [cut]` (the run resource shows at most 300 characters of it); the Output › Codebook entry keeps the whole cause.
 
 **The bundled skill (1.0.47).** In a new VS Code chat that runs in the Copilot harness, where Codebook's MCP tools are not reachable, the skill tells the learner to open a Local chat session, as **Codebook: Set Up AI Chat** says.
 

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-0.2.36.0
+0.2.37.0