refactor(browser): replace workspace with explicit sessions

CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## Unreleased
+
+### ⚠ BREAKING CHANGES
+
+* **browser session model** — replace the browser-facing `--workspace` model with explicit `--session <name>` on `opencli browser *`. Browser commands now require a session name, `browser bind`/`unbind` use `--session`, and bind no longer accepts `--domain`, `--path-prefix`, or `--allow-navigate-bound`. Browser primitives keep their session tab by design; the browser namespace no longer exposes `--keep-tab`. Adapter `--keep-tab` and `browserSession.reuse: 'site'` are unchanged.
+
+### Internal
+
+* **extension 1.0.11** — switch Browser Bridge lease routing from user-facing workspaces to explicit browser sessions.
+
 ## [1.7.16](https://github.com/jackwener/opencli/compare/v1.7.15...v1.7.16) (2026-05-11)
 
 Extension bumped to 1.0.10 (rename adapter-owned tab group `OpenCLI Automation` → `OpenCLI Adapter`). Performance and stability sweep across browser-backed adapters; new external CLI integrations (tg-cli, discord-cli, wx-cli).

README.md

@@ -152,15 +152,15 @@ The agent handles all the `opencli browser` commands internally — you just des
 
 Available browser commands include `open`, `state`, `click`, `type`, `fill`, `select`, `keys`, `wait`, `get`, `find`, `extract`, `frames`, `screenshot`, `scroll`, `back`, `eval`, `network`, `tab list`, `tab new`, `tab select`, `tab close`, `init`, `verify`, and `close`.
 
-`opencli browser open <url>` and `opencli browser tab new [url]` both return a target ID. Use `opencli browser tab list` to inspect the target IDs of tabs that already exist, then pass `--tab <targetId>` to route a command to a specific tab. `tab new` creates a new tab without changing the default browser target; only `tab select <targetId>` promotes that tab to the default target for later untargeted `opencli browser ...` commands.
+`opencli browser` commands require `--session <name>`. `opencli browser --session work open <url>` and `opencli browser --session work tab new [url]` both return a target ID. Use `opencli browser --session work tab list` to inspect target IDs, then pass `--tab <targetId>` to route a command to a specific tab. `tab new` creates a new tab without changing the default browser target; only `tab select <targetId>` promotes that tab to the default target for later untargeted commands in the same session.
 
 ## Core Concepts
 
 ### `browser`: AI Agent browser control
 
 `opencli browser` commands are the low-level primitives that AI Agents use to operate websites. You don't run these manually — instead, install the `opencli-adapter-author` skill into your AI agent, describe what you want in natural language, and the agent handles the browser operations.
 
-For example, tell your agent: *"Help me check my Xiaohongshu notifications"* — the agent will use `opencli browser open`, `state`, `click`, etc. under the hood.
+For example, tell your agent: *"Help me check my Xiaohongshu notifications"* — the agent will use `opencli browser --session <name> open`, `state`, `click`, etc. under the hood.
 
 ### Built-in adapters: stable commands
 
@@ -174,7 +174,7 @@ When the site you need is not yet covered, use the `opencli-adapter-author` skil
 2. Discover the right endpoint — network inspection, initial state, bundle search, token trace, or interceptor fallback.
 3. Decide the auth strategy — `PUBLIC` / `COOKIE` / `INTERCEPT` / `UI` / `LOCAL`.
 4. Decode response fields and design output columns.
-5. `opencli browser analyze <url>` for one-shot recon, then `opencli browser init <site>/<name>` → write adapter → `opencli browser verify <site>/<name>`.
+5. `opencli browser --session recon analyze <url>` for one-shot recon, then `opencli browser --session recon init <site>/<name>` → write adapter → `opencli browser --session recon verify <site>/<name>`.
 6. Persist site knowledge to `~/.opencli/sites/<site>/` so the next adapter for the same site is faster.
 
 ### CLI Hub and desktop adapters
@@ -199,7 +199,7 @@ OpenCLI is not only for websites. It can also:
 | `OPENCLI_DAEMON_PORT` | `19825` | HTTP port for the daemon-extension bridge |
 | `OPENCLI_PROFILE` | — | Browser Bridge profile alias/contextId to use when multiple Chrome profiles are connected |
 | `OPENCLI_WINDOW` | command default | Set to `foreground` or `background` to override Browser Bridge window placement. Browser-backed commands also accept `--window <foreground\|background>`. |
-| `OPENCLI_KEEP_TAB` | command default | Set to `true` or `false` to keep or release the browser tab lease after a browser-backed command. Browser-backed commands also accept `--keep-tab <true\|false>`. |
+| `OPENCLI_KEEP_TAB` | command default | Set to `true` or `false` to keep or release the browser tab lease after a browser-backed adapter command. Browser-backed adapter commands also accept `--keep-tab <true\|false>`. |
 | `OPENCLI_BROWSER_REUSE` | adapter default | Set to `none` or `site` to override adapter browser tab reuse. The `--reuse <none\|site>` flag sets this. |
 | `OPENCLI_BROWSER_CONNECT_TIMEOUT` | `30` | Seconds to wait for browser connection |
 | `OPENCLI_BROWSER_COMMAND_TIMEOUT` | `60` | Seconds to wait for a single browser command |
@@ -208,7 +208,7 @@ OpenCLI is not only for websites. It can also:
 | `OPENCLI_VERBOSE` | `false` | Enable verbose logging (`-v` flag also works) |
 | `DEBUG_SNAPSHOT` | — | Set to `1` for DOM snapshot debug output |
 
-`opencli browser *` uses a foreground browser window and keeps its tab lease by default. Browser-backed adapters use a background automation window and release one-shot tab leases by default. Some interactive adapters default to `--reuse site`, which also keeps the site tab lease; pass `--reuse none` for a one-shot tab.
+`opencli browser *` requires an explicit `--session <name>`, uses a foreground browser window by default, and keeps that session's tab lease until `browser --session <name> close` or idle cleanup. Browser-backed adapters use a background adapter window and release one-shot tab leases by default. Some interactive adapters default to `--reuse site`, which keeps the site tab lease for continuity; pass `--reuse none` for a one-shot tab.
 
 ## Update
 
@@ -409,10 +409,10 @@ See [Plugins Guide](./docs/guide/plugins.md) for creating your own plugin.
 Before writing any adapter code, read the [`opencli-adapter-author` skill](./skills/opencli-adapter-author/SKILL.md). It takes you end-to-end:
 
 - Recon the site and pick a pattern (SPA / SSR / JSONP / Token / Streaming).
-- Discover the right endpoint via `opencli browser network`, `eval`, or the interceptor fallback.
+- Discover the right endpoint via `opencli browser --session <name> network`, `eval`, or the interceptor fallback.
 - Decide auth strategy (`PUBLIC` / `COOKIE` / `INTERCEPT` / `UI` / `LOCAL`).
-- Run `opencli browser analyze <url>` for one-shot recon, decode response fields, design columns, scaffold with `opencli browser init`.
-- Verify with `opencli browser verify <site>/<name>` before shipping.
+- Run `opencli browser --session recon analyze <url>` for one-shot recon, decode response fields, design columns, scaffold with `opencli browser --session recon init`.
+- Verify with `opencli browser --session recon verify <site>/<name>` before shipping.
 
 For long-lived personal commands that should live in your own Git repo, use a local plugin instead; see [Extending OpenCLI](./docs/guide/extending-opencli.md). Quick private adapters can still live at `~/.opencli/clis/<site>/<name>.js`. Site knowledge (endpoints, field maps, fixtures) accumulates in `~/.opencli/sites/<site>/` so the next adapter for the same site starts from context instead of zero.
 

README.zh-CN.md

@@ -136,15 +136,15 @@ Agent 在内部自动处理所有 `opencli browser` 命令——你只需用自
 
 `browser` 可用命令包括：`open`、`state`、`click`、`type`、`fill`、`select`、`keys`、`wait`、`get`、`find`、`extract`、`frames`、`screenshot`、`scroll`、`back`、`eval`、`network`、`tab list`、`tab new`、`tab select`、`tab close`、`init`、`verify`、`close`。
 
-`opencli browser open <url>` 和 `opencli browser tab new [url]` 都会返回 target ID。`opencli browser tab list` 用来查看当前已存在 tab 的 target ID，再通过 `--tab <targetId>` 把命令明确路由到某个 tab。`tab new` 只会新建 tab，不会改变默认浏览器目标；只有显式执行 `tab select <targetId>`，才会把该 tab 设为后续未指定 target 的 `opencli browser ...` 命令的默认目标。
+`opencli browser` 命令必须显式传 `--session <name>`。`opencli browser --session work open <url>` 和 `opencli browser --session work tab new [url]` 都会返回 target ID。`opencli browser --session work tab list` 用来查看当前已存在 tab 的 target ID，再通过 `--tab <targetId>` 把命令明确路由到某个 tab。`tab new` 只会新建 tab，不会改变默认浏览器目标；只有显式执行 `tab select <targetId>`，才会把该 tab 设为同一 session 后续未指定 target 的默认目标。
 
 ## 核心概念
 
 ### `browser`：AI Agent 的浏览器控制层
 
 `opencli browser` 命令是 AI Agent 操作网站的底层原语。你不需要手动运行这些命令——把 `opencli-adapter-author` skill 安装到你的 AI Agent 中，用自然语言描述你想做的事，Agent 会自动处理浏览器操作。
 
-比如你告诉 Agent：*"帮我看看小红书的通知"*——Agent 会在底层调用 `opencli browser open`、`state`、`click` 等命令。
+比如你告诉 Agent：*"帮我看看小红书的通知"*——Agent 会在底层调用 `opencli browser --session <name> open`、`state`、`click` 等命令。
 
 ### 内置适配器：稳定命令
 
@@ -158,7 +158,7 @@ Agent 在内部自动处理所有 `opencli browser` 命令——你只需用自
 2. 发现目标 endpoint——network 精读、initial state、bundle 搜索、token 溯源，或 interceptor 兜底
 3. 定认证策略——`PUBLIC` / `COOKIE` / `INTERCEPT` / `UI` / `LOCAL`
 4. 字段解码 + 设计输出列
-5. `opencli browser analyze <url>` 一步侦察，再 `opencli browser init <site>/<name>` → 写适配器 → `opencli browser verify <site>/<name>`
+5. `opencli browser --session recon analyze <url>` 一步侦察，再 `opencli browser --session recon init <site>/<name>` → 写适配器 → `opencli browser --session recon verify <site>/<name>`
 6. 把站点知识沉到 `~/.opencli/sites/<site>/`，下次写同站点的其他命令直接吃缓存
 
 ### CLI 枢纽与桌面端适配器
@@ -182,15 +182,15 @@ OpenCLI 不只是网站 CLI，还可以：
 |------|--------|------|
 | `OPENCLI_DAEMON_PORT` | `19825` | daemon-extension 通信端口 |
 | `OPENCLI_WINDOW` | 命令默认值 | 设为 `foreground` 或 `background` 来覆盖 Browser Bridge 窗口位置。浏览器型命令也支持 `--window <foreground\|background>` |
-| `OPENCLI_KEEP_TAB` | 命令默认值 | 设为 `true` 或 `false` 来控制浏览器型命令结束后是否保留 tab lease。浏览器型命令也支持 `--keep-tab <true\|false>` |
+| `OPENCLI_KEEP_TAB` | 命令默认值 | 设为 `true` 或 `false` 来控制浏览器型 adapter 命令结束后是否保留 tab lease。浏览器型 adapter 命令也支持 `--keep-tab <true\|false>` |
 | `OPENCLI_BROWSER_CONNECT_TIMEOUT` | `30` | 浏览器连接超时（秒） |
 | `OPENCLI_BROWSER_COMMAND_TIMEOUT` | `60` | 单个浏览器命令超时（秒） |
 | `OPENCLI_CDP_ENDPOINT` | — | Chrome DevTools Protocol 端点，用于远程浏览器或 Electron 应用 |
 | `OPENCLI_CDP_TARGET` | — | 按 URL 子串过滤 CDP target（如 `detail.1688.com`） |
 | `OPENCLI_VERBOSE` | `false` | 启用详细日志（`-v` 也可以） |
 | `DEBUG_SNAPSHOT` | — | 设为 `1` 输出 DOM 快照调试信息 |
 
-`opencli browser *` 默认使用前台窗口并保留 tab lease，直到你手动执行 `opencli browser close` 或等空闲超时。浏览器型 adapter 默认使用后台 automation 窗口并在命令结束后释放 tab lease；如果需要调试最终页面，可以传 `--window foreground --keep-tab true`。
+`opencli browser *` 必须显式传 `--session <name>`，默认使用前台窗口，并保留该 session 的 tab lease，直到你手动执行 `opencli browser --session <name> close` 或等空闲超时。浏览器型 adapter 默认使用后台 adapter 窗口并在命令结束后释放一次性 tab lease；如果需要调试最终页面，可以传 `--window foreground --keep-tab true`。
 
 ## 更新
 
@@ -508,10 +508,10 @@ opencli plugin uninstall my-tool                            # 卸载
 在动代码前，先读 [`opencli-adapter-author` skill](./skills/opencli-adapter-author/SKILL.md)。它把整个流程串起来：
 
 - 侦察站点，选定 pattern（SPA / SSR / JSONP / Token / Streaming）
-- 用 `opencli browser network`、`eval`、interceptor 等找到目标 endpoint
+- 用 `opencli browser --session <name> network`、`eval`、interceptor 等找到目标 endpoint
 - 定认证策略（`PUBLIC` / `COOKIE` / `INTERCEPT` / `UI` / `LOCAL`）
-- 先用 `opencli browser analyze <url>` 一步侦察，再字段解码、设计 columns、`opencli browser init` 生成骨架
-- 交付前用 `opencli browser verify <site>/<name>` 验证
+- 先用 `opencli browser --session recon analyze <url>` 一步侦察，再字段解码、设计 columns、`opencli browser --session recon init` 生成骨架
+- 交付前用 `opencli browser --session recon verify <site>/<name>` 验证
 
 在仓库外写的私有适配器放到 `~/.opencli/clis/<site>/<name>.js`；每个站点的 endpoint、字段映射、抓包样本会累积在 `~/.opencli/sites/<site>/`，下次写同站点的其他命令可以直接复用。
 

cli-manifest.json

@@ -15536,7 +15536,7 @@
     "aliases": [
       "select"
     ],
-    "description": "Open one NotebookLM notebook in the automation workspace by id or URL",
+    "description": "Open one NotebookLM notebook in the adapter session by id or URL",
     "access": "read",
     "domain": "notebooklm.google.com",
     "strategy": "cookie",

clis/notebooklm/current.js

@@ -17,7 +17,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm current', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm current', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const current = await readCurrentNotebooklm(page);
         if (!current) {

clis/notebooklm/get.js

@@ -18,7 +18,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm get', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm get', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const rpcRow = await getNotebooklmDetailViaRpc(page).catch(() => null);
         if (rpcRow)

clis/notebooklm/history.js

@@ -17,7 +17,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm history', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm history', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const rows = await listNotebooklmHistoryViaRpc(page);
         return rows;

clis/notebooklm/note-list.js

@@ -18,7 +18,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm note-list', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm note-list', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const rows = await listNotebooklmNotesFromPage(page);
         if (rows.length > 0)

clis/notebooklm/notes-get.js

@@ -31,7 +31,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm notes-get', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm notes-get', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const query = typeof kwargs.note === 'string' ? kwargs.note : String(kwargs.note ?? '');
         const visible = await readNotebooklmVisibleNoteFromPage(page);

clis/notebooklm/open.js

@@ -7,7 +7,7 @@ cli({
     name: 'open',
     access: 'read',
     aliases: ['select'],
-    description: 'Open one NotebookLM notebook in the automation workspace by id or URL',
+    description: 'Open one NotebookLM notebook in the adapter session by id or URL',
     domain: NOTEBOOKLM_DOMAIN,
     strategy: Strategy.COOKIE,
     browser: true,
@@ -28,7 +28,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new CliError('NOTEBOOKLM_OPEN_FAILED', `NotebookLM notebook "${notebookId}" did not open in the automation workspace`, 'Run `opencli notebooklm list -f json` first and pass a valid notebook id.');
+            throw new CliError('NOTEBOOKLM_OPEN_FAILED', `NotebookLM notebook "${notebookId}" did not open in the adapter session`, 'Run `opencli notebooklm list -f json` first and pass a valid notebook id.');
         }
         if (state.notebookId !== notebookId) {
             console.warn(`[notebooklm open] expected notebook "${notebookId}" but page reports "${state.notebookId}"; continuing`);

clis/notebooklm/open.test.js

@@ -38,7 +38,7 @@ describe('notebooklm open', () => {
             source: 'current-page',
         });
     });
-    it('opens a notebook by id in the automation workspace', async () => {
+    it('opens a notebook by id in the adapter session', async () => {
         const page = {
             goto: vi.fn(async () => { }),
             wait: vi.fn(async () => { }),

clis/notebooklm/source-fulltext.js

@@ -24,7 +24,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm source-fulltext', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm source-fulltext', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const rpcRows = await listNotebooklmSourcesViaRpc(page).catch(() => []);
         const rows = rpcRows.length > 0 ? rpcRows : await listNotebooklmSourcesFromPage(page);

clis/notebooklm/source-get.js

@@ -24,7 +24,7 @@ cli({
         await requireNotebooklmSession(page);
         const state = await getNotebooklmPageState(page);
         if (state.kind !== 'notebook') {
-            throw new EmptyResultError('opencli notebooklm source-get', 'No NotebookLM notebook is open in the automation workspace. Run `opencli notebooklm open <notebook>` first.');
+            throw new EmptyResultError('opencli notebooklm source-get', 'No NotebookLM notebook is open in the adapter session. Run `opencli notebooklm open <notebook>` first.');
         }
         const rpcRows = await listNotebooklmSourcesViaRpc(page).catch(() => []);
         const rows = rpcRows.length > 0 ? rpcRows : await listNotebooklmSourcesFromPage(page);