@omniaibot/win-x64 0.3.11 → 1.1.0

package/bin/omnibot-windows-x64/omnibot/skills/omnibot/references/session-and-tabs.md

@@ -0,0 +1,148 @@
+# Sessions and Tabs
+
+Sessions and tabs are core reliability concepts, not advanced usage.
+
+## Session First
+
+`OMNIBOT_SESSION_TOKEN` is mandatory for every workflow.
+
+```bash
+OMNIBOT_SESSION_TOKEN=research omnibot tabs
+OMNIBOT_SESSION_TOKEN=research omnibot snapshot -i --tab-id <TAB_ID>
+```
+
+The same token keeps one workflow isolated. Different tokens prevent independent agents from sharing default state.
+
+## Tab Explicit
+
+`--tab-id <TAB_ID>` is mandatory for every page-state command.
+
+```bash
+OMNIBOT_SESSION_TOKEN=checkout omnibot get url --tab-id <TAB_ID>
+OMNIBOT_SESSION_TOKEN=checkout omnibot click --tab-id <TAB_ID> @e4
+OMNIBOT_SESSION_TOKEN=checkout omnibot get url --tab-id <TAB_ID>
+```
+
+Do not rely on default tab, active tab, current tab, or previous targeting state.
+
+## Why Both
+
+- Session = workflow isolation.
+- Tab ID = page targeting.
+
+Session alone does not prove which page receives a command. Tab ID alone does not isolate concurrent agents. Use both.
+
+## Ref Scope
+
+`@eN` refs are tab-scoped.
+
+Never reuse `@eN` across tabs. Never assume a ref remains valid after navigation, reload, major DOM changes, or a different snapshot.
+
+```bash
+OMNIBOT_SESSION_TOKEN=checkout omnibot snapshot -i --tab-id <TAB_ID>
+OMNIBOT_SESSION_TOKEN=checkout omnibot click --tab-id <TAB_ID> @e4
+OMNIBOT_SESSION_TOKEN=checkout omnibot snapshot -i --tab-id <TAB_ID>
+```
+
+## Multi-Agent Pattern
+
+Agent A:
+
+```bash
+OMNIBOT_SESSION_TOKEN=research
+```
+
+Agent B:
+
+```bash
+OMNIBOT_SESSION_TOKEN=checkout
+```
+
+Agent C:
+
+```bash
+OMNIBOT_SESSION_TOKEN=debug
+```
+
+Each agent must:
+
+1. Set a stable token.
+2. Discover or create a tab.
+3. Save the tab id.
+4. Pass `--tab-id <TAB_ID>` on every page-state command.
+5. Run `snapshot` -> action -> verify, or `get/is` -> action -> `get/is`.
+
+Example:
+
+```bash
+OMNIBOT_SESSION_TOKEN=research omnibot tab new https://example.com --label research
+OMNIBOT_SESSION_TOKEN=research omnibot get url --tab-id <TAB_ID>
+OMNIBOT_SESSION_TOKEN=research omnibot snapshot -i --tab-id <TAB_ID>
+```
+
+## switch-tab
+
+`switch-tab` is only for casual single-step usage.
+
+Do not build a multi-step workflow on `switch-tab`. It creates implicit state and can be stale or crossed by concurrent agents.
+
+Bad:
+
+```bash
+OMNIBOT_SESSION_TOKEN=research omnibot switch-tab 123
+OMNIBOT_SESSION_TOKEN=research omnibot click @e4
+```
+
+Good:
+
+```bash
+OMNIBOT_SESSION_TOKEN=research omnibot click --tab-id 123 @e4
+OMNIBOT_SESSION_TOKEN=research omnibot snapshot -i --tab-id 123
+```
+
+## focus-tab
+
+Use `focus-tab` only when the user needs to see the browser UI.
+
+It is not a reliable targeting primitive for agent workflows.
+
+```bash
+OMNIBOT_SESSION_TOKEN=demo omnibot focus-tab 123
+OMNIBOT_SESSION_TOKEN=demo omnibot snapshot -i --tab-id 123
+```
+
+## browser claim/release
+
+Use `browser claim` and `browser release` in multi-browser or multi-runtime scenarios where an agent must explicitly own or release a browser tab.
+
+```bash
+OMNIBOT_SESSION_TOKEN=research omnibot browser list
+OMNIBOT_SESSION_TOKEN=research omnibot browser current
+OMNIBOT_SESSION_TOKEN=research omnibot browser claim 123
+OMNIBOT_SESSION_TOKEN=research omnibot snapshot -i --tab-id 123
+OMNIBOT_SESSION_TOKEN=research omnibot browser release 123
+```
+
+Claiming does not replace `--tab-id`. It only documents or coordinates ownership.
+
+## visibility
+
+Visibility controls where automation runs:
+
+- `visible`: controls the user's already-open real browser tabs and preserves visible session state.
+- `background`: avoids foregrounding tabs unless requested.
+- `dedicated-profile`: launches or uses a separate profile.
+- `headless`: uses a headless browser context.
+- `launch`: starts a dedicated-profile or headless browser with an explicit user data dir.
+
+Examples:
+
+```bash
+omnibot visibility status
+omnibot visibility set visible
+omnibot visibility set background
+omnibot visibility set headless --user-data-dir /tmp/omnibot-headless
+omnibot visibility launch headless --user-data-dir /tmp/omnibot-headless
+```
+
+Headless and dedicated-profile modes do not automatically share the user's current tabs or login state. Verify login state before assuming access.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@omniaibot/win-x64",
-  "version": "0.3.11",
+  "version": "1.1.0",
   "description": "omnibot executable for Windows x64",
   "os": [
     "win32"
@@ -8,6 +8,9 @@
   "cpu": [
     "x64"
   ],
+  "bin": {
+    "omnibot": "bin/omnibot-windows-x64/omnibot-windows-x64.exe"
+  },
   "files": [
     "bin/"
   ],
@@ -23,4 +26,4 @@
     "type": "git",
     "url": "git+https://github.com/DennisJcy/omnibot.git"
   }
-}
+}

package/bin/omnibot-win-x64/VERSION

@@ -1 +0,0 @@
-0.3.11

package/bin/omnibot-win-x64/omnibot/skills/omnibot/SKILL.md

@@ -1,159 +0,0 @@
----
-name: omnibot
-description: Use when reading, inspecting, navigating, operating, debugging, installing skills, checking daemon status, or troubleshooting the omnibot CLI extension connection.
-compatibility: Requires omnibot v2 CLI daemon and the omnibot Chromium extension connected to 127.0.0.1:18765.
-allowed-tools: Bash
-metadata:
-  author: omnibot
-  version: "2.0.0"
----
-
-# omnibot CLI
-
-omnibot connects agents to the user's real Chromium browser session through a local daemon and CLI commands.
-
-## Prerequisites
-
-- Confirm the extension is connected with `omnibot doctor` or `omnibot tabs`.
-- For browser actions, run commands through Bash and parse the returned JSON.
-- The CLI auto-starts the local daemon. Users do not need to run `omnibot daemon start` for normal CLI operations.
-
-## Command Discovery
-
-```bash
-omnibot --help
-omnibot scan --help
-omnibot execute-js --help
-omnibot batch --help
-```
-
----
-
-## Install Skills
-
-```bash
-omnibot skills install --agent hermes --profile nuwa
-omnibot skills install --agent opencode
-omnibot skills install --agent claude
-omnibot skills install --agent codex
-```
-
-## Check Runtime
-
-```bash
-omnibot doctor
-omnibot tabs
-```
-
-## Common Failures
-
-- Extension not connected: open Chrome or Edge with the omnibot extension loaded and keep an HTTP/HTTPS tab open.
-- Wrong port: align extension Bridge Port with `--ws-port`; default is `18765`.
-- No tabs: open an HTTP or HTTPS page; extension ignores non-scriptable pages.
-- Daemon startup failed: run `omnibot doctor` and check the daemon log path from the error output.
-
----
-
-## Page Reading
-
-Use real browser rendering when normal HTTP fetches cannot see logged-in, client-rendered, or dynamic content.
-
-```bash
-omnibot scan
-omnibot scan --json
-omnibot full-scan --duration 10
-omnibot full-scan --duration 0 --text-only
-```
-
-**Selection Rules:**
-- Use `scan` before any action because it does not scroll.
-- Use `full-scan` for read-only article/search/feed pages because it scrolls to trigger lazy loading.
-- Use `--json` when controls, forms, links, or list candidates are needed for a follow-up action.
-- Use `--text-only` when the answer needs prose only and HTML tags would waste tokens.
-
-**Search Workflow:**
-```bash
-omnibot navigate 'https://www.google.com/search?q=omnibot+cli'
-omnibot full-scan --duration 5 --text-only
-```
-
----
-
-## Page Operation
-
-Operate pages only after observing the current viewport.
-
-**Required Order:**
-1. Run `omnibot snapshot --json`.
-2. Identify the target control, link, selector, `@eN` reference, or visible text.
-3. Run the smallest operation command.
-4. Run `omnibot snapshot --json` again to verify the page state.
-
-**Common Commands:**
-```bash
-omnibot execute-js "document.querySelector('button[type=submit]').click()"
-omnibot execute-js "window.scrollBy(0, 800)"
-omnibot wait "return !!document.querySelector('.loaded')" --timeout 15 --interval 0.5
-omnibot switch-tab <TAB_ID>
-omnibot focus-tab <TAB_ID>
-omnibot click @e11
-omnibot click @e11 --new-tab
-omnibot dblclick @e11
-omnibot dblclick @e11 --new-tab
-```
-
-**Snapshot and Interaction Order:**
-1. Run `omnibot snapshot --json`.
-2. Find the target `@eN` reference from the snapshot output.
-3. Run `omnibot click @eN`, `omnibot dblclick @eN`, or the matching `--new-tab` variant.
-4. Run `omnibot navigate --same-tab <url>` when only the current tab content should change.
-5. Run `omnibot snapshot --json` again to verify the page changed.
-
-**`@eN` Target Rules:**
-- Prefer `@eN` over raw CSS selectors when snapshot output exposes an accessible role/name pair.
-- Use `@eN` for buttons, links, tabs, tree nodes, list items, and other interactive elements returned by snapshot.
-- Use `--new-tab` when the link or action should open in a new browser tab.
-- Use `dblclick` only for folder/list/tree/text editors and similar controls that need a double activation.
-- Use `navigate --same-tab` for in-page navigation or soft navigations that reuse the current tab.
-- Run `omnibot snapshot --json` after the interaction to confirm the expected DOM/page state change.
-
-**Rules:**
-- Do not use `execute-js` for ordinary page reading.
-- Prefer `wait` over sleeping in shell commands.
-- Use `focus-tab` only when the user needs to see that tab in Chrome.
-- Use `switch-tab` when only the CLI target tab should change.
-
----
-
-## Debugging
-
-Use debug commands when scan output is insufficient or when browser-level APIs are required.
-
-```bash
-omnibot screenshot -o /tmp/omni-shot.png
-omnibot batch --commands-json '[{"cmd":"cdp","method":"DOM.getDocument","params":{"depth":1}}]'
-omnibot execute-js "return location.href"
-```
-
-**Batch Rules:**
-- Use `batch` for multiple CDP or extension commands that share one target tab.
-- Pass `--tab-id` when the target is not the active CLI tab.
-- Keep batch JSON valid and quote it with single quotes in shell commands.
-
----
-
-## Command Reference
-
-| Former MCP tool | CLI command | Default output |
-| --- | --- | --- |
-| `browser_get_tabs` | `omnibot tabs` | JSON |
-| `browser_scan` | `omnibot scan [--json]` | Text |
-| `browser_full_scan` | `omnibot full-scan [--json]` | Text |
-| `browser_execute_js` | `omnibot execute-js <script>` | JSON |
-| `browser_switch_tab` | `omnibot switch-tab <tab_id>` | JSON |
-| `browser_focus_tab` | `omnibot focus-tab <tab_id>` | JSON |
-| `browser_batch` | `omnibot batch --commands-json '<json>'` | JSON |
-| `browser_wait` | `omnibot wait <condition_js>` | JSON |
-| `browser_navigate` | `omnibot navigate <url>` | JSON |
-| `browser_screenshot` | `omnibot screenshot [-o path]` | JSON |
-| `get_license_status` | `omnibot license status` | JSON |