docs(mobile): recorded the chip strip's touch floor and the active-chip measurement The active chip was 73x28 against an idle 32x32, so the shorter target was the one the user was already on. The entry carries that measurement and the `.tap-chip` floor now applied to every chip.
docs(panel-tabs): recorded the touch target sizes and what could not be measured The close control was a 20px box inside a 28px chip; the entry carries the measurement, the coarse-pointer sizes chosen and the limit that the WebKit headless viewport does not report `pointer: coarse`, so only the class contract could be pinned.
docs: listed M24 in the milestone index Also synced the rows the index never gained (19, 20, 22, 23), corrected M13 to partial and the numbering note to name the milestones that actually are planned.
docs(mobile): recorded the phone drawer's own setting and its empty state The phone's right drawer gained a third navigation setting, and the wiki records why it is a separate axis (a drawer has no rail to put beside it), that its open set is the desktop column's own key, that a restored `layout.mobileRightPanel` must not seed a tab or the home page never appears, and that `home` is what stops the seed effect re-opening a tab the user just closed.
docs(shell): documented the draggable chat column width - Recorded the three measurements that shaped the grip: a strip centred on the column's edge is unreachable because the composer card paints above it; the centred column moves twice the pointer delta; and a drag stores nothing until release. - Noted that `custom` has no CSS rule because its number is the user's, written inline, and that a preset must remove that inline value.
docs(workspace-panels): record the collapsed commit history and its row toggle - The modal seeded `selectedHash` from the first commit, so History opened with one commit expanded and a second click could not close it. The page now opens with every row collapsed, one hash in `useCommitExpansion` is the open row (`''` = none), and a row toggles itself.
docs: measured the settings pane band height - The sheet's identity bands are one height and the rule under them is continuous; recorded the 55/64/37–44px measurement, why the modal bar carries the safe-area inset and the inner panes do not, and that `band-height.test.tsx` pins the declared height because happy-dom has no layout.
harnesses: recorded the per-platform lease primitive and its measurement - The lease is `flock` on macOS and an open-file-description lock on Linux since omp 19.0.1; Linux keeps the two lock spaces apart, so a `flock` probe reports a live lease as free (measured in a container: `flock` free, `F_OFD_GETLK` `F_WRLCK`). - Naming the primitive in the panel page too, since it is the file the stats plugin walks.
docs(app-shell): record the grouped Appearance category and its new paths - Appearance is a master/detail category: Palette / Shell / Editor / Editor behaviour, with the group table as data and a `Record` section map that refuses to compile without a renderer. - Two measured consequences recorded: the category belongs in the modal's `p-0` list, and an E2E that wants a control outside the first group must click that group row first. - Settings -> Appearance -> Layout is now Settings -> Appearance -> Shell, so every panel-navigation, surface, rules and tray-icon reference follows the category's own path.
docs(app-shell): recorded the borderless rule set and its measured split - Named the rung-to-role table (`/5`–`/12` separators, `/15` panel and chip edges, `/20` card and input outlines, `/25`+ interaction) that decides what `borderless` clears. - Recorded why the rules clear colour instead of width: `box-sizing: border-box` is global, so a zero width would shorten every box by 2px. - Added the measured proof that this is a paint change: the shell went 11 → 1 visible borders and the settings modal 918 → 19 with every panel rect byte-identical. - Documented the four exemptions, the `outline`-over-`box-shadow` rule for dialogs, the resizer's background-painted hairline, and the specificity needed to beat a plugin's `@scope`d sheet.
navbar visibility: a plugin entry is a control the menu can switch - Recorded that the menu was built from the five built-ins, so an installed plugin's entry (`Stats`) had a row in the navbar and no checkbox; the id is `plugin:<pluginId>` and validation has to accept its shape before the bundle reports in. - Recorded the one-resolver rule and the Chromium measurement against the live `stats` plugin.
docs(panel-tabs): record the rail/tabs choice, the working set and the header policy The wiki gains the reasons the tab style cannot be re-derived from the code: that the choice is two per-column settings defaulting to the rail, that the strip is a working set whose tab id IS the panel id, that tabs are per-session state under `layout.<side>PanelTabs` with `home` stored rather than derived, that a session which never opened a tab lands on the home page instead of the rail's default view, and that a tab-style column suppresses only a header whose every element is the view's name. Each entry carries what was measured (the 40px row with `Usage`'s timestamp and no heading; the persisted `{tabs, active, home}` blob; the drag that reordered by pointer) so a future reader does not re-break it.
docs(workspace): recorded the repo picker portal and label width - the list must be portaled because the right panel clips at 268px, and the trigger no longer caps its label at 130px - measured at a 268px Source Control column in the browser: the 352px list drew outside the panel instead of being clipped
docs(app-shell): the mobile-view switch lives in the sidebar footer - The top navbar's row is now only the right-side toggle plus the shared cluster; the dev-only layout switch moved to the left column's footer and mirrors the phone footer's "Desktop" control in the same slot.
fix(app-shell): the context ring fills its 24-unit grid at a 2.5-unit stroke Recorded the measurements behind the dial's final geometry, replacing the old claim that ContextRing "keeps its own 18px because it is a dial, not a stroke glyph". That claim was wrong twice over: the 18px dial drew a 2px CSS stroke against the lucide icons' 1.33px ("icon context usage di topnavbar lebih besar dari icon di sebelumnya"), and its full-bleed rim had no room to read next to inline glyphs. The page now states the shipped numbers — 24-unit rim / 16.00 CSS px, radius 10.75, 2.5-unit stroke (1.67 CSS px) — and why the rim exceeds the row's own 20 units: a ring is an OUTLINE carrying the whole graphic in one line, measured at 170 unit² of ink against 141 for refresh-cw and 189 for panel-right, where a 20-unit box fell to 114 and read as too small. It also records that getBBox() is not a valid measure here, because it reports fill geometry and ignores the stroke; every figure comes from rasterising the SVG at 20x and taking the alpha bounding box. And that contextRingGeometry takes the OUTER diameter, which is why the constant is the rim rather than the centreline radius.
docs(chat): record the live-run thinking open and the above-viewport anchor rule
docs(workspace-panels): refresh contract, session-state fallback trap, and the refresh glyph copies - A panel's own Refresh must re-read every page it draws, and a write must re-read what it changed: git-forge now reloads list + open detail on Refresh and on a write, command-runner also resyncs its run topic, stats also resyncs the topic after its forced sample. - `useSessionState`'s fallback must be a stable reference: an inline `{}` re-arms the read effect every render and writes a fresh fallback into state — measured at 301 renders and climbing, which hung the Commands panel's first app-level test for 240 s. - A plugin glyph that mirrors a host icon copies lucide's geometry as separate subpaths: merging the four `d`s of `RefreshCw` changed 82 of 147 456 raster pixels at the arrow tips, while the array form measured 0 (lucide 1.47.0).
docs(shells): record the one-local-server rule and the dead-row Remove - The store owns the limit: a Start is refused while a pid this shell recorded still answers, and a recorded pid whose server died does not block (that Start is the restart). - The panel states the running server instead of offering a second button; a dead owned row now offers Remove.
docs(workspace-panels): document the synthesized root row and explicit root pick - Discovery omits '.' for a non-repo root, so the picker synthesizes the row via repoRows(). - resolveRepoForPanel now takes null for "no pick"; an explicit '.' is honored even when the root is not a repo.
docs(harnesses): document OmO over the pi-acp bridge - Added the `omo` transport row: no native ACP server, driven through `pi-acp` with `PI_ACP_PI_COMMAND=omo`. - Recorded the measured surface on 5.1.9 (69 models, 3 thought levels, list/delete) and the fork/transcript facts.
docs(chat): record that every render style shows the card's facts - A style is a frame, not a capability: the generic body, the `Full output` chip and the reminder badge now reach compact, trace, minimal and tui through `hasGenericBody` and `ToolRowActions`. - The trace node's `warn` tone separates an aborted call from a running one. - Measured at 390 and 1440 on the 41-call sample: all five styles render every row with scrollWidth === clientWidth.
docs(sidebar): record why a /skill: turn must count as a message - The 4 KiB scan counted only `"type":"message"` entries, so a chat opened with `/skill:<name>` (a `custom_message` whose content is the expanded SKILL.md) reported `messageCount: 0` and was dropped as header-only. - Measured: 5 chats in one project (0.8-8.6 MB, real titles) missing; the row survived only as the pending `New Session - <timestamp>` row, which expired 30s after the send.
docs(chat): record how a worker's outcome is reported and drawn The measured reasons behind the orchestration fix: omp's flat error fields, the compacted terminal frame, the retry saga that must not report per attempt, the two ACP gaps (an `omp-acp` failure that arrives as ordinary text, and a turn that never settles), and why Kabmo's own three prompts carry an envelope instead of rendering as the operator's bubbles.
docs(build-test): record the fake omp's agentInvoked ack contract - A builtin that opens no turn must be acked `agentInvoked:false`, and that ack must be the only response. The wrapper reads `true` as "this dispatch claimed the turn slot" and holds `promptRunning` until the agent-start watchdog fires, while the `prompt_result {agentInvoked:false}` that follows is suppressed by the run-attribution rule — so nothing clears the flag early. - Measured on release run 44249018348: the fake acked every prompt `true`, the client's `/kabmo-mode` republish left the session reporting busy for ~10 s after the answer landed, and `rename.e2e.ts` was refused with "This session is working" on a session that had finished — 4 of 10 shards red, 1 flake in 6 locally, 0 in 10 after the fix.
docs(omp-engine): record the credential marker the schema generator must read - `orderedSettings()` yields `Setting` instances but the generator destructures the raw `SettingDefinition`, so `definition.isCredential` was always undefined and no row carried `credential: true`. - Names the four rows that now mask and the test that pins them.
docs(chat): record the dispatch-window prompt_result leak - The suppression rule is the whole run, not its streaming half: `streaming` flips at `agent_start`, so a background command answered between the prompt's dispatch and its first frame reached the client and settled the optimistic turn. - Measured on omp 18.22.2: the plan-review `republish` fired on spawn adoption reached the child 43 ms after the prompt and was answered `agentInvoked:false` 160 ms before `agent_start`; 57 of 58 first prompts rendered the user turn twice, 0 of 50 after the guard.
docs(shells): record the child PATH rule for a local server start - The CLI is a `#!/usr/bin/env bun` script (and a Windows shim), so the child's PATH must carry the directories `which` probed. A GUI launch inherits a minimal PATH, so `which` found the CLI in `~/.bun/bin` while the child died with `env: bun: No such file or directory` in `desktop-<port>.log`. - Each shell now widens the child's PATH with the one list its lookup reads; the rules suites assert the widening.
docs(sidebar): record the sort/⋯/timestamp right-rail measurement The action row is outside the list's scroller, so the 8px scrollbar made the rows sit 8px inside the sort icon. The page now states the fix and the measured numbers: `.session-list-gutter` (always-reserved bar, padding = 12px − scrollbar) plus the half-box trailing inset (3px, 4px on touch).
docs(sidebar): recorded the harness filter's placement and forms - The bullet now names the inline position right of Archived and the measured forms behind it, including the 22px Select glyph that keeps the row on one line at the default width. - The old symbol name in the sidebar-chrome bullet is updated to filterRowForms.
docs(harnesses): record a server-spawned ACP child's transcript - Documented the empty-transcript bug measured on a real antigravity worker: `chat_sessions.messages='[]'`, `origin=''`, a parent relation, and the answer only in `session_answers`. - Recorded the two halves: the `AcpTranscriptRecorder` writing `acp_transcripts` and the loader/restore gate reading the local transcript instead of `origin`.