Merge main

packages/vscode/CHANGELOG.md

@@ -1,6 +1,37 @@
-## [Unreleased]
+## [1.21.0] - 2026-08-26
 
+- **Chat context attachments:** diff and file comments, terminal selections, and linked issues/PRs now show in the conversation as compact context cards — source header, captured content behind an expander, your comment below — instead of raw text inside the message.
+- **Chat: comment on a reply.** Select text in a chat message and choose Comment to attach that quote with your note to the next message; the selection stays highlighted while you type.
+- Chat: the view no longer stays stuck on its loading screen on slow or remote connections, including code-server behind a reverse proxy (thanks to @VinciYan).
+- Composer: hovering a context chip above the input opens a stacked preview of everything attached, where comments can be edited in place or items removed before sending.
+- Chat: @ file mentions now rank files and directories together by how well they match, so the file you typed is at the top instead of below unrelated directories. Multi-word queries match in any order, and long paths keep the folder next to the file name visible.
+- Search: Ctrl/Cmd+P now matches the whole file path, not just the file name — searching a folder name finds the files inside it.
+- Search in dropdowns: searchable pickers (agents, models, providers, branches) now put the best matches first, match multi-word queries in any order, and ignore punctuation (so "gpt4o" finds "gpt-4o").
+- Permissions: cards answer to the keyboard with Alt+Enter to allow once, Alt+Shift+Enter to allow always, and Alt+Backspace to deny; the keys are printed on the buttons.
+- Keyboard: dropdown menus and pickers answer Ctrl+N/Ctrl+P for down/up, the session switcher opens focused on your current session, and shortcut labels in tooltips and menus show the binding you actually have set (thanks to @ChangeHow).
+- Chat: Cmd/Ctrl+Shift+T now cycles through every thinking level offered by the selected model instead of skipping levels after reaching the end (thanks to @nimobeeren).
+- Chat: OpenCode notices now share one style.
+- Chat: the timeline dialog now fits small windows instead of squeezing the message list to a couple of rows (thanks to @gaojunran).
+
+## [1.20.0] - 2026-08-23
+
+- **/btw side questions:** type `/btw` followed by your question to ask something off-topic in a temporary session forked from the current conversation. The answer streams into a panel above the composer; collapse it, keep it as a full session, or discard it without touching the chat (thanks to @jaygupta17).
+- **Skills catalog:** browse curated GitHub skill collections in a card-based catalog with cross-source search and direct links to each skill's repository.
+- Settings: the workspace selector on Providers, Agents, MCP, Commands and Skills now only changes what those pages show instead of moving the chat, session list and file tree to another workspace.
+- Settings/Projects: a project can now pin a thinking level next to its model, for models that offer levels.
+- Settings/General: changing the default model, variant or agent no longer repoints an open chat that already carries a model you picked for it. Chats following the default still switch immediately.
+- Settings/Providers: the provider you select no longer jumps to a different one when the chat selection or provider data changes.
+- Settings/Integrations: the experimental page now only lists integrations that can be installed; unavailable and Coming soon entries were removed.
+- Providers: expanded support for custom providers.
+- Sessions created outside OpenChamber now appear in the sidebar and Recent list without a page refresh (thanks to @tomzx).
 - If OpenCode restarts while a response is still running, the chat now stops with an interrupted state and a notification to continue instead of hanging silently (thanks to @sum117).
+- Usage: Z.ai credit limits now appear alongside its other quota windows.
+- Chat: file paths in messages now open from the session's workspace, even if you last browsed files in another workspace (thanks to @tomzx).
+- Chat: app links such as `spotify://` now ask for confirmation before opening another app. You can trust an app link type on one device and manage trusted links in Settings.
+- While a reply streams, the model status line under the last message now turns into the finished message's info row in place, instead of jumping when the reply completes.
+- Chat: newly sent messages and syntax-highlighted code blocks no longer briefly flicker. Bash output can also grow with its content instead of being cut off.
+- Chat: long user messages can be expanded even when their final layout finishes after they first appear.
+- UI: the default dialog close button is easier to click or tap (thanks to @rockinrimmer).
 
 ## [1.19.0] - 2026-08-19
 

packages/vscode/package.json

@@ -2,7 +2,7 @@
   "name": "openchamber",
   "displayName": "OpenChamber",
   "description": "%extension.description%",
-  "version": "1.19.0",
+  "version": "1.21.0",
   "publisher": "fedaykindev",
   "private": true,
   "repository": {
@@ -245,7 +245,7 @@
   },
   "dependencies": {
     "@openchamber/ui": "workspace:*",
-    "@opencode-ai/sdk": "1.18.18",
+    "@opencode-ai/sdk": "1.18.23",
     "adm-zip": "^0.6.0",
     "jsonc-parser": "^3.3.1",
     "react": "^19.1.1",

packages/vscode/src/AgentManagerPanelProvider.ts

@@ -1,4 +1,5 @@
 import * as vscode from 'vscode';
+import { scheduleCachedStateRetries } from './webviewCachedStateRetry';
 import { handleBridgeMessage, type BridgeRequest, type BridgeResponse } from './bridge';
 import { getThemeKindName } from './theme';
 import type { OpenCodeManager, ConnectionStatus } from './opencode';
@@ -23,6 +24,19 @@ export class AgentManagerPanelProvider {
   private _sseStreams = new Map<string, AbortController>();
   private readonly _webviewDevServerUrl: string | null;
 
+  /**
+   * See webviewCachedStateRetry.ts — a single postMessage can be dropped
+   * before the webview bridge is ready, leaving the loading screen stuck.
+   */
+  private _scheduleCachedStateRetries(targetPanel: vscode.WebviewPanel | undefined): void {
+    scheduleCachedStateRetries({
+      target: targetPanel ?? this._panel,
+      getCurrent: () => this._panel,
+      isConnected: () => this._cachedStatus === 'connected',
+      send: () => this._sendCachedState(),
+    });
+  }
+
   constructor(
     private readonly _context: vscode.ExtensionContext,
     private readonly _extensionUri: vscode.Uri,
@@ -64,6 +78,9 @@ export class AgentManagerPanelProvider {
 
     // Send cached connection status
     this._sendCachedState();
+    // The webview bridge may not be ready yet; keep re-sending so a dropped
+    // `connectionStatus` can never leave the webview stuck on its loading screen.
+    this._scheduleCachedStateRetries(this._panel);
 
     // Handle panel disposal
     this._panel.onDidDispose(() => {
@@ -126,6 +143,13 @@ export class AgentManagerPanelProvider {
 
     // Send to webview if it exists
     this._sendCachedState();
+
+    // When we become connected, keep re-sending at staggered delays so the
+    // webview cannot miss the transition (postMessage is dropped if the
+    // webview bridge is not ready yet).
+    if (status === 'connected') {
+      this._scheduleCachedStateRetries(this._panel);
+    }
   }
 
   public notifySettingsSynced(settings: unknown): void {

packages/vscode/src/ChatViewProvider.ts

@@ -1,4 +1,5 @@
 import * as vscode from 'vscode';
+import { scheduleCachedStateRetries } from './webviewCachedStateRetry';
 import { handleBridgeMessage, type BridgeRequest, type BridgeResponse } from './bridge';
 import { getThemeKindName } from './theme';
 import type { OpenCodeManager, ConnectionStatus } from './opencode';
@@ -58,6 +59,19 @@ export class ChatViewProvider implements vscode.WebviewViewProvider {
   private readonly _MESSAGE_TIMEOUT = 5000; // 5 seconds
   private readonly _MAX_RETRIES = 3;
 
+  /**
+   * See webviewCachedStateRetry.ts — a single postMessage can be dropped
+   * before the webview bridge is ready, leaving the loading screen stuck.
+   */
+  private _scheduleCachedStateRetries(targetView: vscode.WebviewView | undefined): void {
+    scheduleCachedStateRetries({
+      target: targetView ?? this._view,
+      getCurrent: () => this._view,
+      isConnected: () => this._cachedStatus === 'connected',
+      send: () => this._sendCachedState(),
+    });
+  }
+
   private _createMessageId(): string {
     return `msg_${Date.now()}_${Math.random().toString(36).slice(2, 8)}`;
   }
@@ -102,6 +116,9 @@ export class ChatViewProvider implements vscode.WebviewViewProvider {
 
     // Send cached connection status and API URL (may have been set before webview was resolved)
     this._sendCachedState();
+    // The webview bridge may not be ready yet; keep re-sending so a dropped
+    // `connectionStatus` can never leave the webview stuck on its loading screen.
+    this._scheduleCachedStateRetries(webviewView);
 
     // Send current active editor file state to the new webview
     this._lastActiveEditorFilePayload = null;
@@ -185,6 +202,13 @@ export class ChatViewProvider implements vscode.WebviewViewProvider {
     
     // Send to webview if it exists
     this._sendCachedState();
+
+    // When we become connected, keep re-sending at staggered delays so the
+    // webview cannot miss the transition (postMessage is dropped if the
+    // webview bridge is not ready yet).
+    if (status === 'connected') {
+      this._scheduleCachedStateRetries(this._view);
+    }
   }
 
   public addTextToInput(text: string) {

packages/vscode/src/DOCUMENTATION.md

@@ -44,6 +44,8 @@ Keep `bridge.ts` as a thin orchestration layer that delegates message handling t
 
 The webview CSP permits `blob:` only for `worker-src` so shared UI parsers can run bounded local decompression off the main thread. Blob scripts remain disallowed by `script-src`.
 
+The webview build emits each worker as one self-contained file. VS Code webviews cannot load workers directly from extension resource URLs or load module imports from inside a worker. The shared Shiki client therefore fetches the built worker, starts it from a `blob:` URL, and relies on the worker CSP allowance above.
+
 - `bridge-localfs-proxy-runtime.ts`
   - Local `/api/fs/read` and `/api/fs/raw` proxy helpers and shared proxy utility helpers.
   - Workspace-contained Markdown gallery images use these local filesystem

packages/vscode/src/SessionEditorPanelProvider.ts

@@ -1,4 +1,5 @@
 import * as vscode from 'vscode';
+import { scheduleCachedStateRetries } from './webviewCachedStateRetry';
 import { handleBridgeMessage, type BridgeRequest, type BridgeResponse } from './bridge';
 import { getThemeKindName } from './theme';
 import type { OpenCodeManager, ConnectionStatus } from './opencode';
@@ -49,6 +50,19 @@ export class SessionEditorPanelProvider {
   private _lastActiveEditorFilePayload: ActiveEditorFilePayload | null = null;
   private readonly _webviewDevServerUrl: string | null;
 
+  /**
+   * See webviewCachedStateRetry.ts — a single postMessage can be dropped
+   * before the webview bridge is ready, leaving the loading screen stuck.
+   */
+  private _scheduleCachedStateRetries(panelId: string, entry: SessionPanelState): void {
+    scheduleCachedStateRetries({
+      target: entry.panel,
+      getCurrent: () => this._panels.get(panelId)?.panel,
+      isConnected: () => this._cachedStatus === 'connected',
+      send: () => this._sendCachedStateToPanel(entry),
+    });
+  }
+
   constructor(
     private readonly _context: vscode.ExtensionContext,
     private readonly _extensionUri: vscode.Uri,
@@ -116,6 +130,9 @@ export class SessionEditorPanelProvider {
 
     void this.updateTheme(vscode.window.activeColorTheme.kind);
     this._sendCachedStateToPanel(state);
+    // The webview bridge may not be ready yet; keep re-sending so a dropped
+    // `connectionStatus` can never leave the webview stuck on its loading screen.
+    this._scheduleCachedStateRetries(panelId, state);
     void this._broadcastActiveEditorFile();
 
     panel.onDidDispose(() => {
@@ -187,6 +204,15 @@ export class SessionEditorPanelProvider {
     for (const entry of this._panels.values()) {
       this._sendCachedStateToPanel(entry);
     }
+
+    // When we become connected, keep re-sending at staggered delays so the
+    // webview cannot miss the transition (postMessage is dropped if the
+    // webview bridge is not ready yet).
+    if (status === 'connected') {
+      for (const [panelId, entry] of this._panels.entries()) {
+        this._scheduleCachedStateRetries(panelId, entry);
+      }
+    }
   }
 
   public notifySettingsSynced(settings: unknown): void {

packages/vscode/src/bridge-settings-runtime.ts

@@ -173,17 +173,20 @@ const readSharedSettingsFromDisk = (): Record<string, unknown> => {
 };
 
 const writeSharedSettingsToDisk = async (changes: Record<string, unknown>): Promise<void> => {
+  let tmp: string | null = null;
   try {
     await fs.promises.mkdir(path.dirname(OPENCHAMBER_SHARED_SETTINGS_PATH), { recursive: true });
     const current = readSharedSettingsFromDisk();
     const next: Record<string, unknown> = { ...current, ...changes };
     // Atomic write: tmp file + rename. Readers never see a partial/truncated
     // JSON that would fail to parse and silently get coerced to {}.
-    const tmp = `${OPENCHAMBER_SHARED_SETTINGS_PATH}.tmp-${process.pid}-${Date.now()}-${Math.random().toString(36).slice(2, 8)}`;
+    tmp = `${OPENCHAMBER_SHARED_SETTINGS_PATH}.tmp-${process.pid}-${Date.now()}-${Math.random().toString(36).slice(2, 8)}`;
     await fs.promises.writeFile(tmp, JSON.stringify(next, null, 2), 'utf8');
     await fs.promises.rename(tmp, OPENCHAMBER_SHARED_SETTINGS_PATH);
   } catch {
-    // ignore
+    if (tmp) {
+      await fs.promises.rm(tmp, { force: true }).catch(() => {});
+    }
   }
 };
 

packages/vscode/src/commandCodeQuota.ts

@@ -1,66 +0,0 @@
-type CommandCodeCredits = {
-  credits?: { monthlyCredits?: number; purchasedCredits?: number; freeCredits?: number };
-  windowLimits?: {
-    fiveHour?: { used?: number; cap?: number; resetAt?: number };
-    weekly?: { used?: number; cap?: number; resetAt?: number };
-  };
-};
-
-type WindowData = { usedPercent: number | null; resetAt: number | null; windowSeconds: number | null; valueLabel: string };
-
-const toWindow = (data: WindowData) => ({
-  usedPercent: data.usedPercent,
-  remainingPercent: data.usedPercent === null ? null : Math.max(0, 100 - data.usedPercent),
-  windowSeconds: data.windowSeconds,
-  resetAfterSeconds: data.resetAt === null ? null : Math.max(0, Math.floor((data.resetAt - Date.now()) / 1000)),
-  resetAt: data.resetAt,
-  resetAtFormatted: null,
-  resetAfterFormatted: null,
-  valueLabel: data.valueLabel,
-});
-
-const isFiniteNumber = (value: unknown): value is number => typeof value === 'number' && Number.isFinite(value);
-const formatCredits = (value: number): string => String(Math.round((value + Number.EPSILON) * 100) / 100);
-
-const parseCredits = (value: unknown): CommandCodeCredits | null => {
-  if (!value || typeof value !== 'object') return null;
-  const payload = value as CommandCodeCredits;
-  return payload;
-};
-
-const parseOrgId = (value: unknown): string | null | undefined => {
-  if (!value || typeof value !== 'object') return undefined;
-  const org = (value as { org?: { id?: unknown } }).org;
-  return typeof org?.id === 'string' && org.id.trim() ? org.id.trim() : null;
-};
-
-const parseCommandCodeCredits = (payload: CommandCodeCredits) => {
-  const windows: Record<string, ReturnType<typeof toWindow>> = {};
-  for (const [label, value] of [['monthly_credits', payload.credits?.monthlyCredits], ['purchased_credits', payload.credits?.purchasedCredits], ['free_credits', payload.credits?.freeCredits]] as const) {
-    if (isFiniteNumber(value)) windows[label] = toWindow({ usedPercent: null, resetAt: null, windowSeconds: null, valueLabel: formatCredits(value) });
-  }
-  for (const [label, limit, seconds] of [['5h', payload.windowLimits?.fiveHour, 5 * 60 * 60], ['weekly', payload.windowLimits?.weekly, 7 * 24 * 60 * 60]] as const) {
-    if (!isFiniteNumber(limit?.used) || !isFiniteNumber(limit.cap) || limit.cap <= 0) continue;
-    const resetAt = isFiniteNumber(limit.resetAt) ? (limit.resetAt < 1_000_000_000_000 ? limit.resetAt * 1000 : limit.resetAt) : null;
-    windows[label] = toWindow({ usedPercent: Math.min(100, Math.max(0, limit.used / limit.cap * 100)), resetAt, windowSeconds: seconds, valueLabel: `${formatCredits(limit.used)} / ${formatCredits(limit.cap)}` });
-  }
-  return windows;
-};
-
-const requestJson = async (path: string, apiKey: string): Promise<unknown> => {
-  const response = await fetch(`https://api.commandcode.ai${path}`, { headers: { Accept: 'application/json', Authorization: `Bearer ${apiKey}` }, signal: AbortSignal.timeout(15_000) });
-  if (response.status === 401 || response.status === 403) throw new Error('Command Code authentication failed');
-  if (!response.ok) throw new Error(`Command Code usage API returned HTTP ${response.status}`);
-  return response.json().catch(() => null);
-};
-
-export const fetchCommandCodeUsage = async (apiKey: string) => {
-  const orgId = parseOrgId(await requestJson('/alpha/whoami', apiKey));
-  if (orgId === undefined) throw new Error('Command Code account could not be determined');
-  const creditsPath = orgId ? `/alpha/billing/credits?orgId=${encodeURIComponent(orgId)}` : '/alpha/billing/credits';
-  const payload = parseCredits(await requestJson(creditsPath, apiKey));
-  if (!payload) throw new Error('Command Code usage data could not be parsed');
-  const windows = parseCommandCodeCredits(payload);
-  if (!Object.keys(windows).length) throw new Error('Command Code usage data could not be parsed');
-  return windows;
-};

packages/vscode/src/opencode.ts

@@ -15,6 +15,17 @@ import { applyProviderEnvAliases } from './provider-env-aliases';
 const t = vscode.l10n.t;
 
 const READY_CHECK_TIMEOUT_MS = 30000;
+
+// Reuse a single output channel across restarts instead of creating (and
+// leaking) a new one on every waitForReady call.
+let managerOutputChannel: vscode.OutputChannel | null = null;
+
+function getManagerOutputChannel(): vscode.OutputChannel {
+  if (!managerOutputChannel) {
+    managerOutputChannel = vscode.window.createOutputChannel('OpenChamberManager');
+  }
+  return managerOutputChannel;
+}
 const WINDOWS_EXECUTABLE_EXTENSIONS = (process.env.PATHEXT || '.EXE;.CMD;.BAT;.COM')
   .split(';')
   .map((ext) => ext.trim().toLowerCase())
@@ -160,6 +171,19 @@ function stripWrappingQuotes(value: string): string {
   return trimmed;
 }
 
+function killProcessTree(pid: number | undefined): void {
+  if (!Number.isInteger(pid)) return;
+  if (process.platform === 'win32') {
+    try {
+      spawnSync('taskkill', ['/PID', String(pid), '/T', '/F'], {
+        stdio: 'ignore', timeout: 5000, windowsHide: true,
+      });
+    } catch {
+      // ignore
+    }
+  }
+}
+
 function appendToPath(dir: string) {
   const trimmed = (dir || '').trim();
   if (!trimmed) return;
@@ -613,7 +637,6 @@ async function waitForReady(
   timeoutMs = 15000,
   authHeaders: Record<string, string> = {}
 ): Promise<ReadyResult> {
-  const outputChannel = vscode.window.createOutputChannel('OpenChamberManager');
   const start = Date.now();
   const candidates = getCandidateBaseUrls(serverUrl);
   let attempts = 0;
@@ -641,7 +664,7 @@ async function waitForReady(
         }
 
         clearTimeout(timeout);
-        outputChannel?.appendLine(
+        getManagerOutputChannel().appendLine(
           `Health check to ${url.toString()} returned ${res.status} with body: ${JSON.stringify(body)}`
         );
 
@@ -743,6 +766,7 @@ async function spawnManagedOpenCodeServer(
   return {
     url,
     close: () => {
+      killProcessTree(child.pid);
       try {
         child.kill('SIGTERM');
       } catch {

packages/vscode/src/quotaProviders.test.ts

@@ -17,7 +17,6 @@ const AUTH = JSON.stringify({
   crof: { key: 'test-token' },
   neuralwatt: { key: 'test-token' },
   'opencode-go': { key: 'test-token' },
-  'command-code': { type: 'oauth', access: 'test-token' },
   'zai-coding-plan': { key: 'test-token' },
   deepseek: { key: 'test-token' },
   anthropic: { access: 'test-token', refresh: 'test-refresh' },
@@ -104,57 +103,6 @@ describe('OpenCode Go quota provider (VS Code parity)', () => {
   });
 });
 
-describe('Command Code quota provider (VS Code parity)', () => {
-  test('uses the OAuth access token and resolves server-backed limits', async () => {
-    const requests: Array<{ url: string; init?: RequestInit }> = [];
-    globalThis.fetch = (async (url: string, init?: RequestInit) => {
-      requests.push({ url, init });
-      return mockResponse(url.endsWith('/alpha/whoami')
-        ? { org: { id: 'org/a' } }
-        : { credits: { monthlyCredits: 120 }, windowLimits: { fiveHour: { used: 25, cap: 100, resetAt: 1_776_000_000 } } });
-    }) as typeof fetch;
-
-    const result = await fetchQuotaForProvider('command-code');
-
-    assert.equal(result.ok, true);
-    assert.deepEqual(requests.map(({ url }) => url), [
-      'https://api.commandcode.ai/alpha/whoami',
-      'https://api.commandcode.ai/alpha/billing/credits?orgId=org%2Fa',
-    ]);
-    assert.equal((requests[0].init?.headers as Record<string, string>).Authorization, 'Bearer test-token');
-    assert.equal(result.usage!.windows['5h']!.usedPercent, 25);
-    assert.equal(result.usage!.windows.monthly_credits!.valueLabel, '120');
-  });
-
-  test('omits orgId for personal accounts', async () => {
-    const urls: string[] = [];
-    globalThis.fetch = (async (url: string) => {
-      urls.push(url);
-      return mockResponse(url.endsWith('/alpha/whoami')
-        ? { user: { id: 'user-1' }, org: null }
-        : { credits: { monthlyCredits: 120 } });
-    }) as typeof fetch;
-
-    const result = await fetchQuotaForProvider('command-code');
-
-    assert.equal(result.ok, true);
-    assert.deepEqual(urls, [
-      'https://api.commandcode.ai/alpha/whoami',
-      'https://api.commandcode.ai/alpha/billing/credits',
-    ]);
-  });
-
-  test('formats fractional credit values for display', async () => {
-    globalThis.fetch = (async (url: string) => mockResponse(url.endsWith('/alpha/whoami')
-      ? { org: null }
-      : { credits: { monthlyCredits: 69.7947070034 }, windowLimits: { fiveHour: { used: 0.2052929966, cap: 14 } } })) as typeof fetch;
-
-    const result = await fetchQuotaForProvider('command-code');
-
-    assert.equal(result.usage!.windows.monthly_credits!.valueLabel, '69.79');
-    assert.equal(result.usage!.windows['5h']!.valueLabel, '0.21 / 14');
-  });
-});
 
 describe('Crof quota provider (VS Code parity)', () => {
   test('reports credits balance as valueLabel with null percent', async () => {
@@ -327,6 +275,33 @@ describe('Z.ai quota provider (VS Code parity)', () => {
     assert.equal(windows['MCP Tools']!.windowSeconds, 30 * 24 * 60 * 60);
     assert.equal(windows['MCP Tools']!.resetAt, 1787128459979);
   });
+
+  test('maps CREDIT_LIMIT entries to windows with credit value labels and plan level', async () => {
+    stubFetchReturning(() => Promise.resolve(mockResponse({
+      code: 200,
+      data: {
+        limits: [
+          { type: 'CREDIT_LIMIT', unit: 3, number: 5, usage: 12000, currentValue: 65, remaining: 11934, percentage: 1, nextResetTime: 1787257978907 },
+          { type: 'CREDIT_LIMIT', unit: 6, number: 1, usage: 60000, currentValue: 65, remaining: 59934, percentage: 1, nextResetTime: 1787844668997 },
+        ],
+        level: 'pro',
+      },
+    })));
+
+    const result = await fetchQuotaForProvider('zai-coding-plan');
+    const windows = result.usage!.windows;
+
+    assert.equal(result.ok, true);
+    assert.equal(result.planLabel, 'pro');
+    assert.equal(windows['5h']!.usedPercent, 1);
+    assert.equal(windows['5h']!.windowSeconds, 5 * 60 * 60);
+    assert.equal(windows['5h']!.resetAt, 1787257978907);
+    assert.equal(windows['5h']!.valueLabel, '65 / 12k credits');
+    assert.equal(windows.weekly!.usedPercent, 1);
+    assert.equal(windows.weekly!.windowSeconds, 7 * 24 * 60 * 60);
+    assert.equal(windows.weekly!.resetAt, 1787844668997);
+    assert.equal(windows.weekly!.valueLabel, '65 / 60k credits');
+  });
 });
 
 describe('NeuralWatt quota provider (VS Code parity)', () => {

packages/vscode/src/quotaProviders.ts

@@ -2,7 +2,6 @@ import fs from 'node:fs';
 import path from 'node:path';
 import os from 'node:os';
 import { fetchOpenCodeGoUsage } from './opencodeGoQuota';
-import { fetchCommandCodeUsage } from './commandCodeQuota';
 import { deleteLegacyOpenCodeGoCredential, readCredential } from './quotaCredentials';
 import { getProviderAuth, updateProviderAuth } from './opencodeAuth';
 
@@ -72,13 +71,31 @@ type ZaiLimit = {
   type?: string;
   number?: number;
   unit?: number;
+  usage?: number;
+  currentValue?: number;
+  remaining?: number;
   nextResetTime?: number;
   percentage?: number;
 };
 
+// CREDIT_LIMIT entries carry `usage` (total credits) and `currentValue` (consumed);
+// TOKENS_LIMIT entries only carry a percentage.
+const formatZaiCreditAmount = (value: number): string => {
+  if (value < 1000) return value.toLocaleString('en-US');
+  return `${Math.round(value / 100) / 10}k`;
+};
+
+const formatZaiCreditValueLabel = (limit: ZaiLimit): string | null => {
+  const used = toNumber(limit.currentValue);
+  const total = toNumber(limit.usage);
+  if (used === null || total === null) return null;
+  return `${formatZaiCreditAmount(used)} / ${formatZaiCreditAmount(total)} credits`;
+};
+
 type ZaiPayload = {
   data?: {
     limits?: ZaiLimit[];
+    level?: string;
   };
 };
 
@@ -411,15 +428,20 @@ const buildResult = (data: {
   configured: boolean;
   usage?: ProviderUsage | null;
   error?: string;
-}): ProviderResult => ({
-  providerId: data.providerId,
-  providerName: data.providerName,
-  ok: data.ok,
-  configured: data.configured,
-  usage: data.usage ?? null,
-  ...(data.error ? { error: data.error } : {}),
-  fetchedAt: Date.now(),
-});
+  planLabel?: string | null;
+}): ProviderResult => {
+  const result: ProviderResult = {
+    providerId: data.providerId,
+    providerName: data.providerName,
+    ok: data.ok,
+    configured: data.configured,
+    usage: data.usage ?? null,
+    ...(data.error ? { error: data.error } : {}),
+    fetchedAt: Date.now(),
+  };
+  if (data.planLabel) result.planLabel = data.planLabel;
+  return result;
+};
 
 const resolveXaiAuth = (): XaiAuthEntry | null => {
   const entry = getProviderAuth('xai');
@@ -750,9 +772,6 @@ export const listConfiguredQuotaProviders = () => {
   const configured = new Set<string>();
   const openCodeGoAuth = normalizeAuthEntry(getAuthEntry(auth, ['opencode-go']));
   if (openCodeGoAuth && (typeof openCodeGoAuth.key === 'string' || typeof openCodeGoAuth.token === 'string')) configured.add('opencode-go');
-  const commandCodeAuth = normalizeAuthEntry(getAuthEntry(auth, ['command-code']));
-  if (commandCodeAuth && (typeof commandCodeAuth.key === 'string' || typeof commandCodeAuth.access === 'string' || typeof commandCodeAuth.token === 'string')) configured.add('command-code');
-  if (process.env.COMMAND_CODE_API_KEY?.trim()) configured.add('command-code');
   if (readCredential('ollama-cloud')) configured.add('ollama-cloud');
   if (readCredential('cursor')) configured.add('cursor');
 
@@ -1291,7 +1310,7 @@ const buildClaudeRateLimitResult = (): ProviderResult => (
         providerName: 'Claude',
         ok: false,
         configured: true,
-        error: 'Rate limited by Anthropic. Retrying shortly.',
+        error: 'Rate limited. Retrying soon.',
       })
 );
 
@@ -2059,16 +2078,19 @@ const fetchZaiQuota = async (): Promise<ProviderResult> => {
     const payload = await response.json() as ZaiPayload;
     const limits = Array.isArray(payload?.data?.limits) ? payload.data.limits : [];
     const windows: Record<string, UsageWindow> = {};
-    for (const tokensLimit of limits.filter((limit) => limit?.type === 'TOKENS_LIMIT')) {
-      const windowSeconds = resolveWindowSeconds(tokensLimit as Record<string, unknown>);
+    // The API renamed TOKENS_LIMIT to CREDIT_LIMIT; field semantics stayed the same,
+    // so both limit types map to the same windows.
+    for (const limit of limits.filter((entry) => entry?.type === 'TOKENS_LIMIT' || entry?.type === 'CREDIT_LIMIT')) {
+      const windowSeconds = resolveWindowSeconds(limit as Record<string, unknown>);
       const windowLabel = resolveWindowLabel(windowSeconds);
-      const resetAt = tokensLimit.nextResetTime ? normalizeTimestamp(tokensLimit.nextResetTime) : null;
-      const usedPercent = typeof tokensLimit.percentage === 'number' ? tokensLimit.percentage : null;
+      const resetAt = limit.nextResetTime ? normalizeTimestamp(limit.nextResetTime) : null;
+      const usedPercent = typeof limit.percentage === 'number' ? limit.percentage : null;
 
       windows[windowLabel] = toUsageWindow({
         usedPercent,
         windowSeconds,
         resetAt,
+        valueLabel: formatZaiCreditValueLabel(limit),
       });
     }
 
@@ -2087,6 +2109,7 @@ const fetchZaiQuota = async (): Promise<ProviderResult> => {
       ok: true,
       configured: true,
       usage: { windows },
+      planLabel: payload?.data?.level || null,
     });
   } catch (error) {
     return buildResult({
@@ -2848,18 +2871,6 @@ const fetchQuotaForProviderUncoalesced = async (providerId: string): Promise<Pro
         return buildResult({ providerId, providerName: 'OpenCode Go', ok: false, configured: true, error: error instanceof Error ? error.message : 'Request failed' });
       }
     }
-    case 'command-code': {
-      try {
-        const entry = normalizeAuthEntry(getAuthEntry(readAuthFile(), ['command-code']));
-        const stored = typeof entry?.key === 'string' ? entry.key : typeof entry?.access === 'string' ? entry.access : typeof entry?.token === 'string' ? entry.token : null;
-        const environment = process.env.COMMAND_CODE_API_KEY?.trim() || null;
-        const apiKey = stored?.trim() || environment;
-        if (!apiKey) return buildResult({ providerId, providerName: 'Command Code', ok: false, configured: false, error: 'Not configured' });
-        return buildResult({ providerId, providerName: 'Command Code', ok: true, configured: true, usage: { windows: await fetchCommandCodeUsage(apiKey) } });
-      } catch (error) {
-        return buildResult({ providerId, providerName: 'Command Code', ok: false, configured: true, error: error instanceof Error ? error.message : 'Request failed' });
-      }
-    }
     case 'cursor':
       return fetchCursorQuota();
     case 'crof':

packages/vscode/src/viteConfig.test.ts

@@ -0,0 +1,11 @@
+import assert from 'node:assert/strict';
+import { readFileSync } from 'node:fs';
+import { describe, test } from 'node:test';
+
+const source = readFileSync(new URL('../vite.config.ts', import.meta.url), 'utf8');
+
+describe('VS Code webview worker build', () => {
+  test('bundles worker imports into one file', () => {
+    assert.match(source, /worker:\s*\{[\s\S]*?inlineDynamicImports:\s*true/);
+  });
+});

packages/vscode/src/webviewCachedStateRetry.ts

@@ -0,0 +1,31 @@
+/**
+ * The webview only leaves its initial loading screen once it receives a
+ * `connectionStatus: connected` message. VS Code drops postMessage calls made
+ * before the webview's acquireVsCodeApi bridge is ready (common in
+ * code-server / slow or flaky networks), so a single send can be lost
+ * forever. Re-sending the cached state at staggered delays bounds the wait
+ * without needing a webview-side ack protocol; the payload is idempotent
+ * (connection status + window focus), so duplicate deliveries are harmless.
+ */
+const CACHED_STATE_RETRY_DELAYS_MS = [500, 1500, 3500, 7000, 12000, 20000];
+
+export function scheduleCachedStateRetries<Target>(input: {
+  /** The panel/view the retries belong to. */
+  target: Target | undefined;
+  /** Reads the provider's CURRENT panel/view, so a replaced target stops its stale retries. */
+  getCurrent: () => Target | undefined;
+  /** Retries only make sense for the connected transition. */
+  isConnected: () => boolean;
+  /** Re-sends the provider's cached state. */
+  send: () => void;
+}): void {
+  if (!input.isConnected()) return;
+  const target = input.target;
+  if (!target) return;
+  for (const delayMs of CACHED_STATE_RETRY_DELAYS_MS) {
+    setTimeout(() => {
+      if (input.getCurrent() !== target) return;
+      input.send();
+    }, delayMs);
+  }
+}

packages/vscode/vite.config.ts

@@ -25,6 +25,13 @@ export default defineConfig(({ mode }) => ({
   },
   worker: {
     format: 'es',
+    // VS Code webviews cannot load module imports from inside a web worker.
+    // Keep the Shiki worker self-contained instead of emitting grammar chunks.
+    rollupOptions: {
+      output: {
+        inlineDynamicImports: true,
+      },
+    },
   },
   define: {
     'process.env.NODE_ENV': JSON.stringify(mode === 'production' ? 'production' : 'development'),