OpenClaw

https://docs.openclaw.ai/tools
Admin
OpenClaw is a self-hosted gateway that connects chat applications like Discord, Telegram, and Slack...
Tokens:386,482
Snippets:2,750
Trust Score:5
Update:1 week ago
Show doc for...
Tokens:
Raw
### /plugins command usage examples

Source: https://docs.openclaw.ai/tools/slash-commands.md

Shows the basic /plugins commands for listing, showing, enabling, disabling, and installing plugins. Install sources include ClawHub, npm, and git. The --force flag is required for arbitrary sources and permits replacement of existing installs.

```text
/plugins
/plugins list
/plugin show context7
/plugins enable context7
/plugins disable context7
/plugins install clawhub:<package>
/plugins install npm:@openclaw/<official-package>
/plugins install npm:<package> --force
/plugins install git:<repository>@<ref> --force
```

--------------------------------

### security.installPolicy JSON5 configuration example

Source: https://docs.openclaw.ai/tools/skills-config.md

Example configuration for the operator install policy. It enables the policy, targets skills and plugins, and runs an external executable with arguments, timeouts, environment variables, and trusted directories. The policy runs after staging and before install/update, and applies to ClawHub skills, uploaded skills, Git/local skills, dependency installers, and plugin sources.

```json5
{
  security: {
    installPolicy: {
      enabled: true,
      // Omit targets to cover every supported target.
      targets: ["skill", "plugin"],
      exec: {
        source: "exec",
        command: "/usr/local/bin/openclaw-install-policy",
        args: ["--json"],
        timeoutMs: 10000,
        noOutputTimeoutMs: 10000,
        maxOutputBytes: 1048576,
        passEnv: ["OPENCLAW_STATE_DIR", "PATH"],
        env: { POLICY_MODE: "strict" },
        trustedDirs: ["/usr/local/bin"],
      },
    },
  },
}
```

--------------------------------

### Example stdin for install policy

Source: https://docs.openclaw.ai/tools/skills-config.md

Shows the JSON input structure that the install policy receives on stdin, including protocolVersion, openclawVersion, targetType, targetName, sourcePath, sourcePathKind, source, origin, request, and skill fields.

```json
{
  "protocolVersion": 1,
  "openclawVersion": "2026.6.1",
  "targetType": "skill",
  "targetName": "weather",
  "sourcePath": "/var/folders/.../openclaw-skill-clawhub/root",
  "sourcePathKind": "directory",
  "source": {
    "kind": "clawhub",
    "authority": "openclaw",
    "mutable": false,
    "network": true
  },
  "origin": {
    "type": "clawhub",
    "registry": "https://clawhub.openclaw.ai",
    "slug": "weather",
    "version": "1.0.0"
  },
  "request": {
    "kind": "skill-install",
    "mode": "install",
    "requestedSpecifier": "clawhub:weather@1.0.0"
  },
  "skill": {
    "installId": "clawhub"
  }
}
```

--------------------------------

### openclaw browser extension setup commands

Source: https://docs.openclaw.ai/tools/chrome-extension.md

Commands to inspect, install, and verify the browser extension setup. The inspect action reads installation state without installing, install prepares automatic local bootstrap, and verify authenticates the local profile relay with the existing per-host key.

```bash
openclaw browser extension setup --action inspect --json
openclaw browser extension setup --action install --json
openclaw browser extension setup --action verify --browser-profile chrome --json
```

--------------------------------

### skills config JSON5 example

Source: https://docs.openclaw.ai/tools/skills-config.md

Example configuration for skills.* in ~/.openclaw/openclaw.json, showing allowBundled, load, install, workshop, and entries settings. Note that built-in image generation should use agents.defaults.mediaModels.image with the core image_generate tool instead of skills.entries.

```json5
{
  skills: {
    allowBundled: ["gemini", "peekaboo"],
    load: {
      extraDirs: ["~/path/to/agent-scripts/skills"],
      allowSymlinkTargets: ["~/path/to/skills"],
      watch: true,
    },
    install: {
      preferBrew: true,
      nodeManager: "npm",
      allowUploadedArchives: false,
    },
    workshop: {
      autonomous: { mode: "auto" },
      maxSkillBytes: 40000,
    },
    entries: {
      "image-lab": {
        enabled: true,
        apiKey: { source: "env", provider: "default", id: "GEMINI_API_KEY" },
        env: { GEMINI_API_KEY: "GEMINI_KEY_HERE" },
      },
      peekaboo: { enabled: true },
      sag: { enabled: false },
    },
  },
}
```

--------------------------------

### Install plugin from marketplace

Source: https://docs.openclaw.ai/tools/plugin.md

Use this command to install a Claude-compatible marketplace plugin.

```bash
openclaw plugins install <plugin> --marketplace <source>
```

--------------------------------

### Install @openclaw/lobster plugin

Source: https://docs.openclaw.ai/tools/lobster.md

Install the official Lobster plugin. Installation applies to a running Gateway automatically; otherwise it takes effect on the next startup.

```bash
openclaw plugins install @openclaw/lobster
```

--------------------------------

### Install the moonshot plugin

Source: https://docs.openclaw.ai/tools/kimi-search.md

Install the @openclaw/moonshot-provider plugin. Installation applies to a running Gateway automatically; otherwise it takes effect on the next startup.

```bash
openclaw plugins install @openclaw/moonshot-provider
```

--------------------------------

### Call quick-index globals and catalog handles

Source: https://docs.openclaw.ai/tools/code-mode/guest-api.md

Shows how to call quick-index globals directly and use callable catalog handles when lookup is needed. The example calls read, searches the catalog, and invokes a search tool.

```javascript
const content = await read({ path: "README.md" });

const [tool] = await catalog.search("...");
const result = await tool({ query: "OpenClaw" });

const [search] = await catalog.search("search the web", { limit: 1 });
const schema = await search.describe();
const hits = await search({ query: "OpenClaw code mode" });
```

--------------------------------

### macOS HTTP: Install and start Speech Swift server

Source: https://docs.openclaw.ai/tools/tts/configuration.md

Install Speech Swift via Homebrew and start its local server on port 8080. Requires Speech Swift v0.0.23 or later; if an older version is installed, run 'brew update && brew upgrade speech' first.

```bash
brew install speech
speech-server --port 8080
```

--------------------------------

### POST /start

Source: https://docs.openclaw.ai/tools/browser-control.md

Starts the browser control server.

```APIDOC
## POST /start

### Description
Starts the browser control server or a browser session.

### Method
POST

### Endpoint
/start

### Query Parameters
- **headless** (boolean) - Optional - Requests a one-shot headless launch for local managed profiles. This override is rejected for attach-only, remote CDP, and existing-session profiles.
- **profile** (string) - Optional - Specifies the profile name.

### Response
#### Success Response (200)
- Confirmation of start.

```

--------------------------------

### tool_describe example: get metadata for mcp:calendar:create_event

Source: https://docs.openclaw.ai/tools/tool-search.md

Example JSON for tool_describe. Accepts the id of a search result and loads full metadata including exact input schema and trusted outputSchema when declared.

```json
{ "id": "mcp:calendar:create_event" }
```

--------------------------------

### /goal command output example

Source: https://docs.openclaw.ai/tools/goal.md

Example output of the /goal command with no arguments, showing the current goal summary including status, objective, tokens used, token budget, and available commands.

```text
Goal
Status: active
Objective: get CI green for PR 87469 and push the fix
Tokens used: 12k
Token budget: 12k/50k

Commands: /goal edit <objective>, /goal pause, /goal complete, /goal clear
```

--------------------------------

### /goal quick start commands

Source: https://docs.openclaw.ai/tools/goal.md

Shows the basic /goal commands: start, view, edit, pause, resume, complete, and clear. The 'start' action is optional; any text after /goal that is not a known action word creates a new goal. Explicit actions like 'start' and 'edit' preserve line breaks, indentation, and repeated spaces inside the objective, while leading and trailing whitespace is trimmed.

```text
/goal start get CI green for PR 87469 and push the fix
/goal
/goal edit get CI green for PR 87469, push the fix, and update docs
/goal pause waiting for CI
/goal resume
/goal complete pushed and verified
/goal clear
```

--------------------------------

### Install OpenClaw browser extension

Source: https://docs.openclaw.ai/tools/chrome-extension.md

Run this command on the machine hosting Chrome. On macOS, it registers the native host and asks Chrome to install the official Store extension. Keep the command running while completing Chrome's setup.

```bash
openclaw browser extension install
```

--------------------------------

### List available providers and models with action=list

Source: https://docs.openclaw.ai/tools/image-generation.md

Use this command to inspect available providers and models at runtime. The action is 'list'.

```text
/tool image_generate action=list
```

--------------------------------

### Set OPENCLAW_TRAJECTORY_FLUSH_TIMEOUT_MS

Source: https://docs.openclaw.ai/tools/trajectory.md

Set the OPENCLAW_TRAJECTORY_FLUSH_TIMEOUT_MS environment variable before starting OpenClaw to increase the cleanup timeout for flushing trajectory rows. The default is 10,000 ms; this example sets it to 30,000 ms for slow disks or large stores.

```bash
export OPENCLAW_TRAJECTORY_FLUSH_TIMEOUT_MS=30000
```

--------------------------------

### POST /trace/start

Source: https://docs.openclaw.ai/tools/browser-control.md

Starts a performance trace.

```APIDOC
## POST /trace/start

### Description
Starts a performance trace.

### Method
POST

### Endpoint
/trace/start

### Query Parameters
- **profile** (string) - Optional - Specifies the profile name.

### Response
#### Success Response (200)
- Confirmation of trace start.

```

--------------------------------

### Linux CLI: Download models and verify synthesis

Source: https://docs.openclaw.ai/tools/tts/configuration.md

Install a speech-core Linux release package, download the ONNX model set once, and verify synthesis before starting OpenClaw. Commands: 'speech download-models' and 'speech speak "Hello from OpenClaw" hello.wav'.

```bash
speech download-models
speech speak "Hello from OpenClaw" hello.wav
```

--------------------------------

### Install OpenClaw and run Gateway and dashboard

Source: https://docs.openclaw.ai/tools

Install OpenClaw on macOS, Linux, or WSL2 with the bash script, or on Windows with the PowerShell command. The installer detects the OS and installs Node if needed. After onboarding, install the Gateway as a background service with 'openclaw gateway install'. Launch the Control UI with 'openclaw dashboard'.

```bash
curl -fsSL https://openclaw.ai/install.sh | bash
```

```powershell
iwr -useb https://openclaw.ai/install.ps1 | iex
```

```bash
openclaw gateway install
```

```bash
openclaw dashboard
```

--------------------------------

### Install Chromium headless shell with repository-pinned Playwright installer

Source: https://docs.openclaw.ai/tools/browser/lightweight.md

Use the repository-pinned Playwright installer to install the Chromium headless shell. The second command prints the selected version, platform download, and install directory. Run from the repository root.

```sh
node node_modules/playwright-core/cli.js install chromium-headless-shell
node node_modules/playwright-core/cli.js install --dry-run chromium-headless-shell
```

--------------------------------

### Install Exa plugin

Source: https://docs.openclaw.ai/tools/exa-search.md

Installs the @openclaw/exa-plugin. Applies to a running Gateway automatically; otherwise takes effect on next startup.

```bash
openclaw plugins install @openclaw/exa-plugin
```

--------------------------------

### Choose xAI OAuth during onboarding

Source: https://docs.openclaw.ai/tools/code-execution.md

Select xAI OAuth during a fresh install to authenticate with device-code flow.

```bash
openclaw onboard --install-daemon --auth-choice xai-oauth
```

--------------------------------

### Configure skills.entries in openclaw.json

Source: https://docs.openclaw.ai/tools/skills.md

Example configuration for enabling, disabling, and configuring skills under skills.entries. The 'image-lab' skill is enabled with an API key from environment variable GEMINI_API_KEY, custom environment variables, and custom config. 'peekaboo' is enabled, 'sag' is disabled. Note: the 'coding-agent' bundled skill is opt-in and requires setting skills.entries.coding-agent.enabled: true and having a supported CLI installed.

```json5
{
  skills: {
    entries: {
      "image-lab": {
        enabled: true,
        apiKey: { source: "env", provider: "default", id: "GEMINI_API_KEY" },
        env: { GEMINI_API_KEY: "GEMINI_KEY_HERE" },
        config: {
          endpoint: "https://example.invalid",
          model: "nano-pro",
        },
      },
      peekaboo: { enabled: true },
      sag: { enabled: false },
    },
  },
}
```

--------------------------------

### Install DuckDuckGo plugin

Source: https://docs.openclaw.ai/tools/duckduckgo-search.md

Installs the @openclaw/duckduckgo-plugin. Installation applies to a running Gateway automatically; otherwise it takes effect on the next startup.

```bash
openclaw plugins install @openclaw/duckduckgo-plugin
```

--------------------------------

### Install plugin from ClawHub

Source: https://docs.openclaw.ai/tools/plugin.md

Use this command to install a plugin from ClawHub, which provides OpenClaw-native discovery, scans, version metadata, and install hints.

```bash
openclaw plugins install clawhub:<package>
```
Tokens:
Raw