{
  "generatedBy": "scripts/write-cli-startup-metadata.ts",
  "generatorSignature": "12bd6b080e140d488134c5233eb6910a0d1257f9",
  "channelOptions": [
    "telegram",
    "whatsapp",
    "discord",
    "irc",
    "googlechat",
    "slack",
    "signal",
    "imessage",
    "feishu",
    "nostr",
    "msteams",
    "mattermost",
    "nextcloud-talk",
    "matrix",
    "raft",
    "line",
    "zalo",
    "clickclack",
    "zalouser",
    "sms",
    "synology-chat",
    "tlon",
    "qa-channel",
    "qqbot",
    "twitch"
  ],
  "channelCatalogSignature": "301287d1e16ccab8dae2ab5db4da898c66b8cdbb",
  "rootHelpBundleSignature": "c38a53f269bb1323a16ecf0985a0ae29b88f0561",
  "browserHelpSourceSignature": "bea2a0ce82cc408e10040bb74f5c86c1ee912649",
  "secretsHelpSourceSignature": "d3111e7b19b45fdb6cacbf8afc652e38afb02bda",
  "nodesHelpSourceSignature": "198481884d659ad9ddd9dd75a040c1599f33f1dc",
  "subcommandHelpSourceSignature": "a0988fe2be16daf60092d7a5fa50beaa19b8ba00",
  "browserHelpText": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw browser [options] [command]\n\nManage OpenClaw's dedicated browser (Chrome/Chromium)\n\nOptions:\n  --browser-profile <name>  Browser profile name (default from config)\n  --expect-final            Wait for final response (agent) (default: false)\n  -h, --help                Display help for command\n  --json                    Output machine-readable JSON (default: false)\n  --timeout <ms>            Timeout in ms (default: \"30000\")\n  --token <token>           Gateway token (if required)\n  --url <url>               Gateway WebSocket URL (defaults to\n                            gateway.remote.url when configured)\n\nCommands:\n  click                     Click an element by ref from snapshot\n  click-coords              Click viewport coordinates\n  close                     Close a tab (tab reference optional)\n  console                   Get recent console messages\n  cookies                   Read/write cookies\n  create-profile            Create a new browser profile\n  delete-profile            Delete a browser profile\n  dialog                    Arm the next modal dialog (alert/confirm/prompt)\n  doctor                    Check browser plugin readiness\n  download                  Click a ref and save the resulting download\n  drag                      Drag from one ref to another\n  errors                    Get recent page errors\n  evaluate                  Evaluate a function against the page or a ref\n  extension                 Chrome extension load path and pairing\n  fill                      Fill a form with JSON field descriptors\n  focus                     Focus a tab by tab reference\n  highlight                 Highlight an element by ref\n  hover                     Hover an element by ai ref\n  navigate                  Navigate the current tab to a URL\n  open                      Open a URL in a new tab\n  pdf                       Save page as PDF\n  press                     Press a key\n  profiles                  List all browser profiles\n  requests                  Get recent network requests (best-effort)\n  reset-profile             Reset browser profile (moves it to Trash)\n  resize                    Resize the viewport\n  responsebody              Wait for a network response and return its body\n  screenshot                Capture a screenshot (prints the saved path)\n  scrollintoview            Scroll an element into view by ref from snapshot\n  select                    Select option(s) in a select element\n  set                       Browser environment settings\n  snapshot                  Capture a snapshot (default: ai; aria is the\n                            accessibility tree)\n  start                     Start the browser (no-op if already running)\n  status                    Show browser status\n  stop                      Stop the browser (best-effort)\n  storage                   Read/write localStorage/sessionStorage\n  tab                       Tab shortcuts (index-based)\n  tabs                      List open tabs\n  trace                     Record a Playwright trace\n  type                      Type into an element by ref from snapshot\n  upload                    Arm file upload for the next file chooser\n  wait                      Wait for time, selector, URL, load state, or JS\n                            conditions\n  waitfordownload           Wait for the next download (and save it)\n\nExamples:\n  openclaw browser status\n  openclaw browser start\n  openclaw browser start --headless\n  openclaw browser stop\n  openclaw browser tabs\n  openclaw browser open https://example.com\n  openclaw browser focus abcd1234\n  openclaw browser close abcd1234\n  openclaw browser screenshot\n  openclaw browser screenshot --full-page\n  openclaw browser screenshot --ref 12\n  openclaw browser snapshot\n  openclaw browser snapshot --format aria --limit 200\n  openclaw browser snapshot --efficient\n  openclaw browser snapshot --labels\n  openclaw browser navigate https://example.com\n  openclaw browser resize 1280 720\n  openclaw browser click 12 --double\n  openclaw browser click-coords 120 340\n  openclaw browser type 23 \"hello\" --submit\n  openclaw browser press Enter\n  openclaw browser hover 44\n  openclaw browser drag 10 11\n  openclaw browser select 9 OptionA OptionB\n  openclaw browser upload /tmp/openclaw/uploads/file.pdf\n  openclaw browser upload media://inbound/file.pdf\n  openclaw browser fill --fields '[{\"ref\":\"1\",\"value\":\"Ada\"}]'\n  openclaw browser dialog --accept\n  openclaw browser wait --text \"Done\"\n  openclaw browser evaluate --fn '(el) => el.textContent' --ref 7\n  openclaw browser evaluate --fn 'const title = document.title; return title;'\n  openclaw browser console --level error\n  openclaw browser pdf\n\nDocs: https://docs.openclaw.ai/cli/browser\n\n",
  "secretsHelpText": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw secrets [options] [command]\n\nSecrets runtime controls\n\nOptions:\n  -h, --help  Display help for command\n\nCommands:\n  apply       Apply a previously generated secrets plan\n  audit       Audit plaintext secrets, unresolved refs, and precedence drift\n  configure   Interactive secrets helper (provider setup + SecretRef mapping +\n              preflight)\n  help        Display help for command\n  reload      Re-resolve secret references and atomically swap runtime snapshot\n\nDocs: https://docs.openclaw.ai/gateway/security\n\n",
  "nodesHelpText": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw nodes [options] [command]\n\nManage gateway-owned nodes (pairing, status, invoke, and media)\n\nOptions:\n  -h, --help  Display help for command\n\nCommands:\n  approve     Approve a pending pairing request\n  camera      Capture camera media from a paired node\n  canvas      Capture or render canvas content from a paired node\n  describe    Describe a node (capabilities + supported invoke commands)\n  help        Display help for command\n  invoke      Invoke a command on a paired node\n  list        List pending and paired nodes\n  location    Fetch location from a paired node\n  notify      Send a local notification on a node (mac only)\n  pending     List pending pairing requests\n  push        Send an APNs test push to an iOS node\n  reject      Reject a pending pairing request\n  remove      Remove a paired node entry\n  rename      Rename a paired node (display name override)\n  screen      Capture screen recordings from a paired node\n  status      List known nodes with connection status and capabilities\n\nExamples:\n  openclaw nodes status\n    List known nodes with live status.\n  openclaw nodes pairing pending\n    Show pending node pairing requests.\n  openclaw nodes remove --node <id|name|ip>\n    Remove a stale paired node entry.\n  openclaw nodes invoke --node <id> --command system.which --params '{\"name\":\"uname\"}'\n    Invoke a node command directly.\n  openclaw nodes camera snap --node <id>\n    Capture a photo from a node camera.\n\nDocs: https://docs.openclaw.ai/cli/nodes\n\n",
  "subcommandHelpText": {
    "doctor": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw doctor [options]\n\nHealth checks + quick fixes for the gateway and channels\n\nOptions:\n  --all                       With --lint: run all registered checks, including\n                              opt-in checks (default: false)\n  --allow-exec                Allow doctor to execute exec SecretRefs while\n                              verifying configured secrets (default: false)\n  --deep                      Scan system services for extra gateway installs\n                              (default: false)\n  --fix                       Apply recommended repairs (alias for --repair)\n                              (default: false)\n  --force                     Apply aggressive repairs (overwrites custom\n                              service config) (default: false)\n  --generate-gateway-token    Generate and configure a gateway token (default:\n                              false)\n  -h, --help                  Display help for command\n  --json                      With --lint or --post-upgrade: emit\n                              machine-readable JSON output (default: false)\n  --lint                      Run read-only health checks and report findings\n                              (default: false)\n  --no-workspace-suggestions  Disable workspace memory system suggestions\n  --non-interactive           Run without prompts (safe migrations only)\n                              (default: false)\n  --only <id>                 With --lint: run only the specified check id\n                              (repeatable) (default: [])\n  --post-upgrade              Emit plugin-compat findings only (machine-readable\n                              with --json) (default: false)\n  --repair                    Apply recommended repairs without prompting\n                              (default: false)\n  --severity-min <level>      With --lint: drop findings below this severity\n                              (info|warning|error)\n  --skip <id>                 With --lint: skip a specific check id (repeatable)\n                              (default: [])\n  --yes                       Accept defaults without prompting (default: false)\n\nDocs: https://docs.openclaw.ai/cli/doctor\n\n",
    "gateway": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw gateway [options] [command]\n\nRun, inspect, and query the WebSocket Gateway\n\nOptions:\n  --allow-unconfigured       Allow gateway start without enforcing\n                             gateway.mode=local in config (does not repair\n                             config) (default: false)\n  --auth <mode>              Gateway auth mode\n                             (\"none\"|\"token\"|\"password\"|\"trusted-proxy\")\n  --bind <mode>              Bind mode\n                             (\"loopback\"|\"lan\"|\"tailnet\"|\"auto\"|\"custom\").\n                             Defaults to config gateway.bind (or loopback).\n  --claude-cli-logs          Deprecated alias for --cli-backend-logs (default:\n                             false)\n  --cli-backend-logs         Only show CLI backend logs in the console (includes\n                             stdout/stderr) (default: false)\n  --compact                  Alias for \"--ws-log compact\" (default: false)\n  --dev                      Create a dev config + workspace if missing (no\n                             BOOTSTRAP.md) (default: false)\n  --force                    Kill any existing listener on the target port\n                             before starting (default: false)\n  -h, --help                 Display help for command\n  --password <password>      Password for auth mode=password\n  --password-file <path>     Read gateway password from file\n  --port <port>              Port for the gateway WebSocket\n  --raw-stream               Log raw model stream events to jsonl (default:\n                             false)\n  --raw-stream-path <path>   Raw stream jsonl path\n  --reset                    Reset dev config + credentials + sessions +\n                             workspace (requires --dev) (default: false)\n  --tailscale <mode>         Tailscale exposure mode (\"off\"|\"serve\"|\"funnel\")\n  --tailscale-reset-on-exit  Reset Tailscale serve/funnel configuration on\n                             shutdown (default: false)\n  --token <token>            Shared token required in connect.params.auth.token\n                             (default: OPENCLAW_GATEWAY_TOKEN env if set)\n  --verbose                  Verbose logging to stdout/stderr (default: false)\n  --ws-log <style>           WebSocket log style (\"auto\"|\"full\"|\"compact\")\n                             (default: \"auto\")\n\nCommands:\n  call                       Call a Gateway method\n  diagnostics                Export local support diagnostics\n  discover                   Discover gateways via Bonjour (local + wide-area if\n                             configured)\n  health                     Fetch Gateway health\n  install                    Install the Gateway service\n                             (launchd/systemd/schtasks)\n  probe                      Show gateway reachability, auth capability, and\n                             read-probe summary (local + remote)\n  restart                    Restart the Gateway service\n                             (launchd/systemd/schtasks)\n  run                        Run the WebSocket Gateway (foreground)\n  stability                  Fetch payload-free Gateway stability diagnostics\n  start                      Start the Gateway service\n                             (launchd/systemd/schtasks)\n  status                     Show gateway service status + probe\n                             connectivity/capability\n  stop                       Stop the Gateway service (launchd/systemd/schtasks)\n  uninstall                  Uninstall the Gateway service\n                             (launchd/systemd/schtasks)\n  usage-cost                 Fetch usage cost summary from session logs\n\nExamples:\n  openclaw gateway run\n    Run the gateway in the foreground.\n  openclaw gateway status\n    Show service status plus connectivity/capability.\n  openclaw gateway discover\n    Find local and wide-area gateway beacons.\n  openclaw gateway stability\n    Show recent stability diagnostics.\n  openclaw gateway call health\n    Call a gateway RPC method directly.\n\nDocs: https://docs.openclaw.ai/cli/gateway\n\n",
    "models": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw models [options] [command]\n\nModel discovery, scanning, and configuration\n\nOptions:\n  --agent <id>     Agent id to inspect (overrides OPENCLAW_AGENT_DIR)\n  -h, --help       Display help for command\n  --status-json    Output JSON (alias for `models status --json`) (default:\n                   false)\n  --status-plain   Plain output (alias for `models status --plain`) (default:\n                   false)\n\nCommands:\n  aliases          Manage model aliases\n  auth             Manage model auth profiles\n  fallbacks        Manage model fallback list\n  image-fallbacks  Manage image model fallback list\n  list             List models (configured by default)\n  scan             Scan OpenRouter free models for tools + images\n  set              Set the default model\n  set-image        Set the image model\n  status           Show configured model state\n\nDocs: https://docs.openclaw.ai/cli/models\n\n",
    "plugins": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw plugins [options] [command]\n\nManage OpenClaw plugins and extensions\n\nOptions:\n  -h, --help   Display help for command\n\nCommands:\n  build        Generate simple tool plugin metadata\n  disable      Disable a plugin in config\n  doctor       Report plugin load issues\n  enable       Enable a plugin in config\n  init         Create a plugin project\n  inspect      Inspect plugin details\n  install      Install a plugin or hook pack (path, archive, npm spec, git repo,\n               clawhub:package, or marketplace entry)\n  list         List discovered plugins\n  marketplace  Inspect Claude-compatible plugin marketplaces\n  registry     Inspect or rebuild the persisted plugin registry\n  search       Search ClawHub plugin packages\n  uninstall    Uninstall a plugin\n  update       Update installed plugins and tracked hook packs\n  validate     Validate simple tool plugin metadata\n\nDocs: https://docs.openclaw.ai/cli/plugins\n\n",
    "sessions": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw sessions [options] [command]\n\nList stored conversation sessions\n\nOptions:\n  --active <minutes>  Only show sessions updated within the past N minutes\n  --agent <id>        Agent id to inspect (default: configured default agent)\n  --all-agents        Aggregate sessions across all configured agents (default:\n                      false)\n  -h, --help          Display help for command\n  --json              Output as JSON (default: false)\n  --limit <count>     Max sessions to show (default: 100; use \"all\" for full\n                      output)\n  --store <path>      Path to session store (default: resolved from config)\n  --verbose           Verbose logging (default: false)\n\nCommands:\n  cleanup             Run session-store maintenance now\n  compact             Compact a stored session transcript via the running\n                      gateway\n  export-trajectory   Export a redacted trajectory bundle for a stored session\n  list                List stored conversation sessions\n  tail                Tail human-readable session trajectory progress\n\nExamples:\n  openclaw sessions\n    List all sessions.\n  openclaw sessions --agent work\n    List sessions for one agent.\n  openclaw sessions --all-agents\n    Aggregate sessions across agents.\n  openclaw sessions --active 120\n    Only last 2 hours.\n  openclaw sessions --limit 25\n    Show the newest 25 sessions.\n  openclaw sessions --json\n    Machine-readable output.\n  openclaw sessions --store ./tmp/sessions.json\n    Use a specific session store.\n\nShows token usage per session when the agent reports it; set agents.defaults.contextTokens to cap the window and show %.\n\nDocs: https://docs.openclaw.ai/cli/sessions\n\n",
    "tasks": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw tasks [options] [command]\n\nInspect durable background tasks and TaskFlow state\n\nOptions:\n  -h, --help        Display help for command\n  --json            Output as JSON (default: false)\n  --runtime <name>  Filter by kind (subagent, acp, cron, cli)\n  --status <name>   Filter by status (queued, running, succeeded, failed,\n                    timed_out, cancelled, lost)\n\nCommands:\n  audit             Show stale or broken background tasks and TaskFlows\n  cancel            Cancel a running background task\n  flow              Inspect durable TaskFlow state under tasks\n  list              List tracked background tasks\n  maintenance       Preview or apply tasks and TaskFlow maintenance\n  notify            Set task notify policy\n  show              Show one background task by task id, run id, or session key\n"
  },
  "rootHelpText": "\nOpenClaw 2026.7.1-2 (0790d9f) — All your chats, one OpenClaw.\n\nUsage: openclaw [options] [command]\n\nOptions:\n  --container <name>   Run the CLI inside a running Podman/Docker container\n                       named <name> (default: env OPENCLAW_CONTAINER)\n  --dev                Dev profile: isolate state under ~/.openclaw-dev, default\n                       gateway port 19001, and shift derived ports\n                       (browser/canvas)\n  -h, --help           Display help for command\n  --log-level <level>  Global log level override for file + console\n                       (silent|fatal|error|warn|info|debug|trace)\n  --no-color           Disable ANSI colors\n  --profile <name>     Use a named profile (isolates\n                       OPENCLAW_STATE_DIR/OPENCLAW_CONFIG_PATH under\n                       ~/.openclaw-<name>)\n  -V, --version        output the version number\n\nCommands:\n  Hint: commands suffixed with * have subcommands. Run <command> --help for details.\n  acp *                Run an ACP bridge backed by the Gateway\n  agent                Run an agent turn via the Gateway (use --local for\n                       embedded)\n  agents *             Manage isolated agents (workspaces + auth + routing)\n  approvals *          Manage exec approvals (gateway or node host)\n  attach               Attach Claude Code to a gateway session with scoped MCP\n                       tools\n  audit                Inspect metadata-only agent run and tool action records\n  backup *             Create and verify local backup archives for OpenClaw\n                       state\n  capability *         Run provider capability commands (fallback alias: infer)\n  channels *           Manage connected chat channels and accounts\n  chat                 Open a local terminal UI (alias for tui --local)\n  clawbot *            Legacy clawbot command aliases\n  commitments *        List and manage inferred follow-up commitments\n  completion           Generate shell completion script\n  config *             Non-interactive config helpers\n                       (get/set/patch/unset/file/schema/validate). Run without\n                       subcommand for guided setup.\n  configure            Interactive configuration for credentials, channels,\n                       gateway, and agent defaults\n  crestodian           Open the ring-zero setup and repair helper\n  cron *               Manage cron jobs (via Gateway)\n  daemon *             Manage the Gateway service (launchd/systemd/schtasks)\n  dashboard            Open the Control UI with your current token\n  devices *            Device pairing and auth tokens\n  directory *          Lookup contact and group IDs (self, peers, groups) for\n                       supported chat channels\n  dns *                DNS helpers for wide-area discovery (Tailscale + CoreDNS)\n  docs                 Search the live OpenClaw docs\n  doctor               Health checks + quick fixes for the gateway and channels\n  exec-approvals *     Manage exec approvals (alias for approvals)\n  exec-policy *        Show or synchronize requested exec policy with host\n                       approvals\n  gateway *            Run, inspect, and query the WebSocket Gateway\n  health               Fetch health from the running gateway\n  help                 Display help for command\n  hooks *              Manage internal agent hooks\n  infer *              Run provider-backed inference commands through a stable\n                       CLI surface\n  logs                 Tail gateway file logs via RPC\n  mcp *                Manage OpenClaw mcp.servers config and channel bridge\n  memory *             Search, inspect, and reindex memory files\n  message *            Send, read, and manage messages and channel actions\n  migrate *            Import state from another agent system\n  models *             Model discovery, scanning, and configuration\n  node *               Run and manage the headless node host service\n  nodes *              Manage gateway-owned nodes (pairing, status, invoke, and\n                       media)\n  onboard              Guided setup for auth, models, Gateway, workspace,\n                       channels, and skills\n  pairing *            Secure DM pairing (approve inbound requests)\n  plugins *            Manage OpenClaw plugins and extensions\n  promos *             Discover and claim promotional model offers from ClawHub\n  proxy *              Run the OpenClaw debug proxy and inspect captured traffic\n  qr                   Generate a mobile pairing QR code and setup code\n  reset                Reset local config/state (keeps the CLI installed)\n  sandbox *            Manage sandbox containers (Docker-based agent isolation)\n  secrets *            Secrets runtime controls\n  security *           Audit local config and state for common security\n                       foot-guns\n  sessions *           List stored conversation sessions\n  setup                Alias for openclaw onboard\n  skills *             List and inspect available skills\n  status               Show channel health and recent session recipients\n  system *             System tools (events, heartbeat, presence)\n  tasks *              Inspect durable background tasks and TaskFlow state\n  terminal             Open a local terminal UI (alias for tui --local)\n  transcripts *        Inspect stored transcripts\n  tui                  Open a terminal UI connected to the Gateway\n  uninstall            Uninstall the gateway service + local data (CLI remains)\n  update *             Update OpenClaw and inspect update channel status\n  webhooks *           Webhook helpers and integrations\n  worktrees *          Create, inspect, restore, and clean up managed worktrees\n\nExamples:\n  openclaw onboard\n    Run guided setup for a local Gateway, workspace, auth, and channels.\n  openclaw setup\n    Create the baseline config, workspace, and session folders.\n  openclaw configure\n    Change models, Gateway, channels, plugins, skills, and health checks.\n  openclaw status\n    Check Gateway, channel, model, and recent-session status.\n  openclaw doctor --fix\n    Repair common config, service, plugin, and channel problems.\n  openclaw channels add\n    Add or update a chat channel account with guided prompts.\n  openclaw channels status\n    See connected messaging accounts and login state.\n  openclaw --dev gateway\n    Run a dev Gateway (isolated state/config) on ws://127.0.0.1:19001.\n  openclaw gateway run --force\n    Start the Gateway and replace anything bound to its port.\n  openclaw models status\n    Show model/provider auth health before running agents.\n  openclaw plugins list\n    Inspect enabled, disabled, and installed plugins.\n  openclaw agent --to +15555550123 --message \"Run summary\" --deliver\n    Run one agent turn through the Gateway and optionally deliver the reply.\n  openclaw message send --channel telegram --target @mychat --message \"Hi\"\n    Send via your Telegram bot.\n\nDocs: https://docs.openclaw.ai/cli\n\n"
}