SwarmClaw

https://github.com/swarmclawai/swarmclaw
Admin
SwarmClaw is an open-source, self-hosted AI agent runtime and multi-agent framework for building...
Tokens:38,368
Snippets:596
Trust Score:7.2
License:MIT
Update:2 months ago
Show doc for...
Tokens:
Raw
### Local Development Setup

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CONTRIBUTING.md

Steps to clone the repository, install dependencies, and start the development server.

```bash
git clone https://github.com/swarmclawai/swarmclaw.git
cd swarmclaw
npm install
npm run dev      # starts the dev server on port 3456
```

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

### Run SwarmClaw from repository

Source: https://github.com/swarmclawai/swarmclaw/blob/main/README.md

Clone the repository and start the application using the provided quickstart script.

```bash
git clone https://github.com/swarmclawai/swarmclaw.git
cd swarmclaw
nvm use
npm run quickstart
```

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

### Install Node Package

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/resourceful-problem-solving/SKILL.md

Install necessary Node.js packages for scripting.

```bash
npm install <package>
```

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

### Start Development Server

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

This command starts the development server. Never stop or restart this process without user confirmation, as it may be actively in use by developers or other agents.

```bash
npm run dev
```

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

### Install SwarmClaw via npm

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.5.68-launch-plan.md

Installs the SwarmClaw CLI globally using npm. Ensure you have Node.js and npm installed.

```bash
npm install -g @swarmclawai/swarmclaw@1.5.68
```

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

### Install SwarmClaw CLI

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/swarmclaw/SKILL.md

Install the SwarmClaw command-line interface globally using npm.

```bash
npm install -g swarmclaw
```

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

### SwarmClaw Tool Usage Example

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Example of how to structure a gws command within SwarmClaw's tool arguments, using 'drive files list' as a case.

```json
{"args":["drive","files","list"],"params":{"pageSize":5}}
```

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

### Install Python Package

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/resourceful-problem-solving/SKILL.md

Install necessary Python packages for scripting.

```bash
pip install <package>
```

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

### Define Setup Defaults

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

This file serves as the single source of truth for provider defaults, starter agent kits, and default model selections. Add new provider or agent preset defaults here.

```typescript
import "src/lib/setup-defaults.ts";
```

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

### Summarize Content from URL, File, or YouTube

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/summarize/SKILL.md

Basic usage examples for summarizing different types of content. Specify the content source and optionally a model for summarization.

```bash
summarize "https://example.com" --model google/gemini-3-flash-preview
summarize "/path/to/file.pdf" --model google/gemini-3-flash-preview
summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto
```

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

### Execute SQLite Query

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

This is an example of executing a SQLite query within the sandbox environment.

```bash
sqlite3 :memory: 'SELECT 1+1'
```

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

### Check gws Installation and Health

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Run this command to verify that the gws CLI is installed correctly and is functioning as expected.

```bash
gws doctor
```

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

### Install and Run Pi Coding Agent

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/coding-agent/SKILL.md

Install the Pi coding agent using npm and run it for coding tasks. It supports non-interactive mode and different providers/models.

```bash
# Install: npm install -g @mariozechner/pi-coding-agent
cd /path/to/project && pi 'Your task'
```

```bash
# Non-interactive mode
pi -p 'Summarize src/'
```

```bash
# Different provider/model
pi --provider openai --model gpt-4o-mini -p 'Your task'
```

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

### Get Help for gws Commands

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Inspect the available commands and their usage for the gws CLI or specific resources like Drive.

```bash
gws help
gws drive help
gws drive files help
```

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

### Run System Diagnostics

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Check the system's health status. Start with the quick health summary, then run full diagnostics if needed.

```bash
swarmclaw system status --json
```

```bash
swarmclaw setup doctor --json
```

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

### Install SwarmClaw CLI

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Install the SwarmClaw CLI globally using npm. This command is required before using any other SwarmClaw CLI commands.

```bash
npm i -g @swarmclawai/swarmclaw
```

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

### Execute JavaScript Script

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

This is an example of executing a JavaScript script within the sandbox environment.

```bash
js-exec 'console.log("hello")'
```

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

### Install SwarmClaw Skill

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw-skill.md

Installs the SwarmClaw skill into your OpenClaw environment. This command is used to add the skill's functionality to your agent.

```bash
openclaw skills install swarmclaw
```

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

### Execute Python Script

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

This is an example of executing a Python script within the sandbox environment.

```bash
python3 -c 'print("hello")'
```

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

### Skill File YAML Frontmatter Example

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/skills.md

Skills can include optional YAML frontmatter to declare metadata such as the skill name, description, and requirements like binary dependencies.

```yaml
---
name: my-skill
description: What this skill teaches
metadata:
  openclaw:
    requires:
      bins: ["gh"]
---
```

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

### Search for CLI Tools

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/resourceful-problem-solving/SKILL.md

Check if a suitable CLI tool is already installed or available in package managers.

```bash
which <tool>
```

```bash
brew search <tool>
```

```bash
pip search <tool>
```

```bash
npm search <tool>
```

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

### API Call with Credential Injection

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

This example demonstrates making an API call to OpenAI, injecting an API key from an environment variable for authentication, and parsing the JSON response.

```bash
curl -s -H "Authorization: Bearer $OPENAI_API_KEY" \
  https://api.openai.com/v1/models | jq '.data[].id'
```

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

### Authenticate and Verify GitHub CLI

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Authenticate your GitHub CLI installation for one-time use. Verify the authentication status.

```bash
gh auth login
```

```bash
gh auth status
```

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

### Query GitHub API for Repository Statistics

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Make an API query to get repository statistics like star count and fork count. Replace `owner/repo` with the repository path.

```bash
gh api repos/owner/repo --jq '{stars: .stargazers_count, forks: .forks_count}'
```

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

### Get a Google Doc

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Retrieve the content of a specific Google Document using its ID.

```bash
gws docs get --document-id <DOC_ID>
```

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

### Get Goal Details

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieves detailed information about a specific goal using its ID. Useful for understanding the status and specifics of a goal.

```bash
swarmclaw goals get <goalId> --json
```

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

### Get a Google Sheet

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Retrieve information about a specific Google Sheet using its spreadsheet ID.

```bash
gws sheets spreadsheets get --spreadsheet-id <SPREADSHEET_ID>
```

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

### Get Task Details

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieve details for a specific task using its ID. Use `--json` for structured output.

```bash
swarmclaw tasks get <taskId> --json
```

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

### Get a Google Drive File

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Retrieve details of a specific file from Google Drive using its file ID.

```bash
gws drive files get --file-id <FILE_ID>
```

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

### Pull SwarmClaw Docker Image

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.5.68-launch-plan.md

Pulls the SwarmClaw Docker image from GHCR. This requires Docker to be installed and configured.

```bash
ghcr.io/swarmclawai/swarmclaw:v1.5.68
```

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

### Get Knowledge Source Detail

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieves full details of a knowledge source, including metadata and indexed chunks. Use the source ID to specify the source.

```bash
swarmclaw knowledge source-get <sourceId> --json
```

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

### Get Agent Details

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieve full details for a specific agent using its ID. Use `--json` for structured output.

```bash
swarmclaw agents get <agentId> --json
```

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

### Update Schedule

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Updates an existing schedule, for example, to enable or disable it. Use the schedule ID to target the update.

```bash
swarmclaw schedules update <scheduleId> --data '{"enabled":false}' --json
```

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

### Get Task Board Metrics

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieve task board metrics for a specified range (24h, 7d, or 30d). Use `--json` for structured output.

```bash
swarmclaw tasks metrics --json --query range=7d
```

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

### Short Social Post Example

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.5.68-launch-plan.md

A concise social media post announcing SwarmClaw v1.5.68. Ideal for platforms like X or LinkedIn, including a link to the website.

```text
SwarmClaw v1.5.68 is out.

It is a self-hosted runtime for autonomous AI agent teams: missions, memory, MCP tools, schedules, delegation, desktop installers, and 23 providers.

New: Launch Week Growth Sprint template, plus fresh docs and Next.js security updates.
https://swarmclaw.ai
```

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

### Initialize New Skill Directory

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/skill-creator/SKILL.md

Use the provided init script to create the basic directory structure for a new skill. Specify the skill name and the output path, optionally including resource directories.

```bash
python3 {baseDir}/scripts/init_skill.py <skill-name> --path <output-directory> [--resources scripts,references,assets] [--examples]
```

```bash
python3 {baseDir}/scripts/init_skill.py my-skill --path skills
```

```bash
python3 {baseDir}/scripts/init_skill.py my-skill --path skills --resources scripts,references
```

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

### Loading Skills with the Skills Tool

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/swarmclaw.md

Demonstrates how to list all available skills, read a specific skill's documentation, or search for skills related to a query using the 'skills' tool.

```json
{
  "tool": "skills",
  "action": "list"
}
```

```json
{
  "tool": "skills",
  "action": "read",
  "name": "tools/files"
}
```

```json
{
  "tool": "skills",
  "action": "search",
  "query": "github pr"
}
```

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

### Set Up Goal Hierarchy and Link Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Create organizational goals and associate agent tasks with them using goal IDs. You will need task and agent IDs.

```bash
swarmclaw goals create --data '{"title":"Increase Q2 revenue 20%","level":"organization","description":"Hit $1.2M ARR by end of Q2"}' --json
```

```bash
# Use the returned goalId to link agent tasks
```

```bash
swarmclaw tasks list --json --query agentId=<sales-agent-id>
```

```bash
swarmclaw tasks update <taskId> --data '{"goalId":"<goalId>"}' --json
```

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

### Get Live Agent Status

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieve the live status of an agent, including running chats and current tasks. Use `--json` for structured output.

```bash
swarmclaw agents status <agentId> --json
```

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

### Load Skills using the Skills Tool

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/swarmclaw/SKILL.md

Demonstrates how to load skills using the 'skills' tool. Use 'list' to see available skills, 'read' to load a specific skill's documentation, and 'search' to find skills related to a query.

```json
{ "tool": "skills", "action": "list" }
```

```json
{ "tool": "skills", "action": "read", "name": "tools/files" }
```

```json
{ "tool": "skills", "action": "search", "query": "github pr" }
```

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

### Build Website

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.5.68-launch-plan.md

Builds the project's website using npm. This command is used to generate the static site for deployment.

```bash
npm run build
```

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

### Create a Chatroom

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Create a new chatroom with a specified name and description. Provide chatroom details in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw chatrooms create --data '{"name":"Strategy Room","description":"Multi-agent planning"}' --json
```

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

### Upload File as Knowledge

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Uploads a local file to be indexed as a knowledge source. The system will process the file content for agent use.

```bash
swarmclaw knowledge upload ./docs/handbook.pdf --json
```

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

### List All Projects

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Retrieve a list of all available projects. This action requires no parameters.

```json
{ "action": "projects.list" }
```

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

### Get a Gmail Message

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

Retrieve a specific email message from Gmail using its message ID.

```bash
gws gmail users messages get --user-id me --message-id <MESSAGE_ID>
```

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

### Run OpenCode

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/coding-agent/SKILL.md

Execute tasks using the OpenCode agent. Ensure you are in the project directory before running.

```bash
cd /path/to/project && opencode run 'Your task'
```

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

### Create a New PDF with nano-pdf

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/nano-pdf/SKILL.md

Use this command to generate a new PDF file from scratch, providing a natural-language description of the desired content. The tool will create a single-page summary based on the prompt.

```bash
nano-pdf create output.pdf "Create a one-page summary of quarterly results with a header, bullet points, and a footer"
```

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

### Create an Agent

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Create a new agent by providing its configuration details in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw agents create --data '{"name":"Research Bot","provider":"anthropic","model":"claude-sonnet-4-20250514"}' --json
```

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

### Create a New Task

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Use this action to create a new task with a title, description, and priority. Ensure all required parameters are provided.

```json
{
  "action": "tasks.create",
  "title": "Review PR #55",
  "description": "Check for type safety issues and test coverage",
  "priority": "high"
}
```

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

### List Knowledge Sources

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Lists all knowledge sources with their summaries. Provides an overview of available knowledge bases.

```bash
swarmclaw knowledge sources --json
```

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

### Create and Manage Multi-Agent Chatrooms

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Initiate a chatroom for multiple agents to collaborate. You will need the IDs of the agents you wish to add.

```bash
swarmclaw agents list --json
```

```bash
# Identify the strategy and research agent IDs
```

```bash
swarmclaw chatrooms create --data '{"name":"Q2 Strategy Brainstorm"}' --json
```

```bash
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<strategy-agent-id>"}' --json
```

```bash
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<research-agent-id>"}' --json
```

```bash
swarmclaw chatrooms chat <chatroomId> --data '{"message":"Brainstorm our Q2 growth strategy. Consider market trends, competitor moves, and our current strengths."}' --json
```

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

### Get Chat Message History

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Retrieve the message history for a specific chat session. Use `--json` for structured output.

```bash
swarmclaw chats messages <chatId> --json
```

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

### Build Electron Application

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Build the Electron application for the maintainer's operating system. This is part of the verification gate for release.

```bash
npm run electron:build
```

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

### Run Unit Tests

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Execute all unit tests for the project. Ensure this passes before proceeding with the release.

```bash
npm test
```

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

### Create Goal

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Creates a new goal with a title, level, and description. Supports hierarchical goal management.

```bash
swarmclaw goals create --data '{"title":"Increase revenue 20%","level":"organization","description":"Q2 revenue target"}' --json
```

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

### Useful npm Scripts for Development

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CONTRIBUTING.md

Common npm scripts for linting, testing, and building the project during local development.

```bash
npm run lint:baseline           # release gate; must pass before merging
```

```bash
npx tsx --test path/to/file.test.ts   # run a single Node-test file
```

```bash
npm run test:cli                # CLI test suite
```

```bash
npm run test:openclaw           # OpenClaw integration suite
```

```bash
npm run test:setup              # setup wizard suite
```

```bash
NODE_ENV=production npm run build:ci   # production build sanity check
```

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

### Create and Assign a Task

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Create a new task with a title, description, and assigned agent ID. Provide task details in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw tasks create --data '{"title":"Analyse competitor pricing","description":"Research and compare competitor pricing strategies","agentId":"<agentId>"}' --json
```

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

### Build for CI

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Execute the CI build process. This command prepares the project for continuous integration checks.

```bash
npm run build:ci
```

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

### Deploy SwarmClaw with Docker

Source: https://github.com/swarmclawai/swarmclaw/blob/main/README.md

Build and run the application container using Docker Compose.

```bash
git clone https://github.com/swarmclawai/swarmclaw.git
cd swarmclaw
mkdir -p data
touch .env.local
docker compose up -d --build
```

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

### List Open Issues

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

List all open issues for a specified repository. Replace `owner/repo` with the repository path.

```bash
gh issue list --repo owner/repo --state open
```

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

### List Available Skills

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/skills.md

Use the 'list' action to browse all available skills. This returns their names and short descriptions.

```json
{ "action": "list" }
```

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

### List All Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

List all tasks in the task board. Supports filtering by status, agent, and project. Use `--json` for structured output.

```bash
swarmclaw tasks list --json
```

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

### Search Skills by Keyword

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/skills.md

Use the 'search' action with a query to find skills. This searches skill file names, descriptions, and content for keyword matches and returns ranked results.

```json
{ "action": "search", "query": "browser screenshot" }
```

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

### List Chatrooms

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

List all available chatrooms. Use `--json` for structured output.

```bash
swarmclaw chatrooms list --json
```

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

### Query GitHub API to List Labels

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Make an API query to list all labels in a repository. Replace `owner/repo` with the repository path.

```bash
gh api repos/owner/repo/labels --jq '.[].name'
```

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

### Configure Default Model via JSON

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/summarize/SKILL.md

Set a default model for summarization using a local configuration file. This avoids needing to specify the model with each command.

```json
{
  "model": "openai/gpt-5.2"
}
```

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

### List Schedules

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Lists all configured schedules. This is a foundational command for managing automated tasks.

```bash
swarmclaw schedules list --json
```

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

### Create Manual Knowledge Entry

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Creates a new knowledge entry from manual text content, including title, content, kind, and tags. Useful for adding structured information.

```bash
swarmclaw knowledge create --data '{"title":"Pricing Policy","content":"Standard pricing is...","kind":"manual","tags":["pricing"]}' --json
```

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

### Configure OpenTelemetry OTLP Export

Source: https://github.com/swarmclawai/swarmclaw/blob/main/README.md

Set these environment variables to enable OTLP trace export for SwarmClaw.

```bash
OTEL_ENABLED=true
OTEL_SERVICE_NAME=swarmclaw
OTEL_EXPORTER_OTLP_ENDPOINT=https://your-collector:4318
OTEL_EXPORTER_OTLP_HEADERS=Authorization=Bearer your-token
```

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

### Navigate to a URL

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/browser.md

Use the navigate action to go to a specific web page. This is the first step in most browser interactions.

```json
{ "action": "navigate", "url": "https://example.com/dashboard" }
```

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

### List Google Drive Files

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/google-workspace/SKILL.md

List files in Google Drive with a specified page size. Use this to find file IDs.

```bash
gws drive files list --params '{"pageSize":10}'
```

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

### Run End-to-End Tests

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Execute end-to-end tests to verify the application's behavior in a production-like environment. This is crucial for release verification.

```bash
npm run test:e2e
```

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

### List Agents, Chats, and In-Progress Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Check the current status of agents, active chats, and tasks that are currently in progress.

```bash
swarmclaw agents list --json
```

```bash
swarmclaw chats list --json
```

```bash
swarmclaw tasks list --json --query status=in_progress
```

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

### Create a New Chat Session

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Create a new chat session for a specified agent. Provide agent ID in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw chats create --data '{"agentId":"<agentId>"}' --json
```

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

### Save Script as Extension

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/resourceful-problem-solving/SKILL.md

Offer to save a successful script as a reusable extension using the `extension_creator` tool.

```bash
extension_creator <script_name>
```

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

### Show HN Intro Draft

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.5.68-launch-plan.md

A draft for a Hacker News 'Show HN' post, detailing SwarmClaw's technical aspects and new release features. Focus on technical details and solicit feedback.

```text
I built SwarmClaw because I wanted a local control plane for autonomous agent teams instead of a pile of one-off chats and scripts.

It supports missions with budgets/reports, MCP servers, memory, schedules, chat connectors, desktop installers, and 23 providers including CLI backends like Claude Code, Codex, OpenCode, Gemini, Cursor, Qwen, Goose, plus OpenAI-compatible APIs.

The newest release adds a Launch Week Growth Sprint template, mostly because I wanted the product to help organize its own release. Feedback on the install path, mission model, and MCP/tooling model would be especially useful.
```

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

### List All Agents

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Retrieve a list of all available agents. This action requires no parameters and is useful for discovering agent capabilities.

```json
{ "action": "agents.list" }
```

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

### Check for Existing Skills

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/resourceful-problem-solving/SKILL.md

Use `use_skill` to search for a matching skill for common tasks.

```bash
use_skill <skill_name>
```

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

### Sync Knowledge Source

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Re-syncs a file or URL knowledge source to ensure its content is up-to-date. Use the source ID to target the sync.

```bash
swarmclaw knowledge source-sync <sourceId> --json
```

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

### Create an Issue

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Create a new issue with a specified title and body. Customize the title and body as needed.

```bash
gh issue create --title "Bug: something broken" --body "Details..."
```

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

### Read a Specific Skill

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/skills.md

Use the 'read' action to load the full content of a skill file by its name. Skill names use a path-style notation.

```json
{ "action": "read", "name": "tools/files" }
```

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

### Write File Content

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/files.md

Creates a new file or overwrites an existing one with the provided content. Parent directories are created automatically.

```json
{ "action": "write", "path": "src/config.ts", "content": "export const PORT = 3000\n" }
```

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

### Review PRs by Cloning to Temp Folder

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/coding-agent/SKILL.md

Clone a repository to a temporary folder for safe PR reviews. This prevents modifications within the SwarmClaw project directory. After cloning, checkout the specific PR and initiate the review.

```bash
# Clone to temp for safe review
REVIEW_DIR=$(mktemp -d)
git clone https://github.com/user/repo.git $REVIEW_DIR
cd $REVIEW_DIR && gh pr checkout 130
codex review --base origin/main
```

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

### Skill Directory Structure

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/skill-creator/SKILL.md

Illustrates the standard file and directory layout for a SwarmClaw skill, including required and optional components.

```markdown
skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter (name + description, required)
│   └── Markdown instructions (required)
└── Bundled Resources (optional)
    ├── scripts/      — Executable code (Python/Bash/etc.)
    ├── references/   — Documentation loaded into context as needed
    └── assets/       — Files used in output (templates, icons, fonts)
```

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

### List Goals

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Lists all hierarchical goals across different levels (organization, team, project, agent, task). Provides an overview of objectives.

```bash
swarmclaw goals list --json
```

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

### Run Lint Baseline Check

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

Execute this command before any production release to ensure no new linting issues are introduced. If violations are fixed, use the update command to lock in improvements.

```bash
npm run lint:baseline
```

```bash
npm run lint:baseline:update
```

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

### Use HintTip Component for Contextual Help

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

The `HintTip` component, represented by a '?' icon, provides contextual help for fields that may not be self-explanatory. It's recommended for any non-obvious UI elements.

```typescript
import { HintTip } from "src/components/shared/hint-tip.tsx";
```

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

### Multi-Image Composition

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/nano-banana-pro/SKILL.md

Combine multiple images into a single scene based on a text prompt. Use the `-i` flag for each input image, supporting up to 14 images.

```bash
uv run {baseDir}/scripts/generate_image.py --prompt "combine these into one scene" --filename "output.png" -i img1.png -i img2.png -i img3.png
```

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

### Multi-Step Login and Data Extraction

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/browser.md

A sequence of browser actions to log in to a web application, wait for the dashboard to load, and then extract specific metrics. This demonstrates a typical multi-step interaction flow.

```json
{ "action": "navigate", "url": "https://app.example.com/login" }
```

```json
{ "action": "type", "selector": "#email", "text": "user@example.com" }
```

```json
{ "action": "type", "selector": "#password", "text": "$APP_PASSWORD" }
```

```json
{ "action": "click", "text": "Log In" }
```

```json
{ "action": "wait", "selector": ".dashboard-loaded" }
```

```json
{ "action": "extract_text", "selector": ".metrics-panel" }
```

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

### Store a New Memory

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/memory.md

Use the 'store' action to save a new piece of information. Provide a title, value, and category for the memory.

```json
{
  "action": "store",
  "title": "User prefers dark mode",
  "value": "The user explicitly asked for dark mode in all UI components. Use dark backgrounds (#1a1a2e) with light text (#e0e0e0).",
  "category": "preference"
}
```

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

### Search Memories by Query

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/memory.md

Use the 'search' action to find relevant memories. The query parameter is essential for matching. Supports semantic-style matching.

```json
{ "action": "search", "query": "database schema preferences" }
```

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

### Create a New Task

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Use this command to create a new task for an agent. Ensure you have the agent's ID.

```bash
swarmclaw tasks create --data '{"title":"Analyse competitor pricing","description":"Research and compare competitor pricing strategies, identify gaps and opportunities","agentId":"<research-agent-id>"}' --json
```

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

### Triage Open Issues with Details

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

List open issues with their number, title, labels, and creation date, formatted for triage. Replace `owner/repo` with the repository path.

```bash
gh issue list --repo owner/repo --state open --json number,title,labels,createdAt \
  --jq '.[] | "[\(.number)] \(.title) - \([.labels[].name] | join(", ")) (\(.createdAt[:10]))"'
```

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

### List Pull Requests with GitHub CLI

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

List pull requests for a specified repository. Ensure to replace `owner/repo` with the actual repository path.

```bash
gh pr list --repo owner/repo
```

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

### Validate Skill Structure

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/skill-creator/SKILL.md

Use this command to validate the structure and frontmatter of your skill folder. Ensure the path to your skill folder is correctly provided.

```bash
python3 {baseDir}/scripts/quick_validate.py <path/to/skill-folder>
```

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

### Import GitHub Issues as Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Import GitHub issues as tasks, specifying the repository, labels, and the agent to assign them to. Provide import details in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw tasks import-github --data '{"repo":"owner/repo","labels":["bug"],"agentId":"<id>"}' --json
```

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

### Create Schedule

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Creates a new schedule for an agent to perform a task at a specified cron time. Requires agent ID, cron expression, and a message.

```bash
swarmclaw schedules create --data '{"name":"Daily Report","agentId":"<id>","cron":"0 9 * * *","message":"Generate the daily summary report"}' --json
```

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

### Generate an Image

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/nano-banana-pro/SKILL.md

Use this command to generate a new image based on a text prompt. Specify the output filename and desired resolution.

```bash
uv run {baseDir}/scripts/generate_image.py --prompt "your image description" --filename "output.png" --resolution 1K
```

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

### Create a Pull Request

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Create a new pull request with a specified title and body. Customize the title and body as needed.

```bash
gh pr create --title "feat: add feature" --body "Description"
```

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

### Persistent File Write (Host Mode)

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

This snippet demonstrates a persistent file write operation, which requires the host execution mode to be enabled.

```json
{ "code": "echo 'hello' > output.txt", "persistent": true }
```

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

### Knowledge Hygiene Summary

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Provides a summary of knowledge hygiene, highlighting stale, duplicate, or broken sources. Useful for identifying maintenance needs.

```bash
swarmclaw knowledge hygiene --json
```

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

### List Knowledge Entries

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Lists all knowledge entries, which can be searched using query parameters. This is the primary way to view managed knowledge.

```bash
swarmclaw knowledge list --json
```

```bash
swarmclaw knowledge list --json --query q=pricing
```

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

### List Chats

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

List all chat sessions. Use `--json` for structured output.

```bash
swarmclaw chats list --json
```

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

### Generate Electron App Icons

Source: https://github.com/swarmclawai/swarmclaw/blob/main/resources/README.md

Run this command to regenerate app icons from the source image. This script is macOS-specific and requires Xcode command line tools.

```bash
node scripts/gen-icons.mjs
```

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

### List Open Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Retrieve a list of tasks that are currently open. This action can be filtered by status.

```json
{ "action": "tasks.list", "status": "open" }
```

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

### Explore SwarmClaw CLI Commands

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Discover available command groups and detailed help for specific commands and flags within the SwarmClaw CLI.

```bash
swarmclaw --help                    # List all command groups
```

```bash
swarmclaw <group> --help            # List commands in a group
```

```bash
swarmclaw <group> <action> --help   # Show flags for a specific command
```

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

### Generate Images with DALL-E 2

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/openai-image-gen/SKILL.md

Generate images using the DALL-E 2 model, specifying size and count. DALL-E 2 only supports 'standard' quality.

```bash
python3 {baseDir}/scripts/gen.py --model dall-e-2 --size 512x512 --count 4
```

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

### List Issues with JSON Output and Filtering

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

List issues, outputting only the number and title, then format the output to show 'number: title'. Replace `owner/repo` with the repository path.

```bash
gh issue list --repo owner/repo --json number,title --jq '.[] | "\(.number): \(.title)"'
```

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

### Add Member to Chatroom

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Adds an agent to a specified chatroom. Use this to include agents in group conversations.

```bash
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<agent1>"}' --json
```

```bash
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<agent2>"}' --json
```

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

### Generate Multiple Images with GPT

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/openai-image-gen/SKILL.md

Use the GPT image model to generate a specified number of images. Supports various options for prompt, count, and output directory.

```bash
python3 {baseDir}/scripts/gen.py --count 16 --model gpt-image-1
```

```bash
python3 {baseDir}/scripts/gen.py --prompt "ultra-detailed studio photo of a lobster astronaut" --count 4
```

```bash
python3 {baseDir}/scripts/gen.py --size 1536x1024 --quality high --out-dir ./out/images
```

```bash
python3 {baseDir}/scripts/gen.py --model gpt-image-1.5 --background transparent --output-format webp
```

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

### Take a Full Page Screenshot

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/browser.md

Capture a screenshot of the entire web page, including content not currently in view. Use the `fullPage` parameter set to true.

```json
{ "action": "screenshot", "fullPage": true }
```

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

### List Directory Contents

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/files.md

Lists directory entries, including file types. Use 'depth' to control the level of recursion.

```json
{ "action": "list", "path": "src/components" }
```

```json
{ "action": "list", "path": "src", "depth": 2 }
```

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

### View a Specific Workflow Run

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

View details of a specific workflow run. Replace `<run-id>` with the run ID and `owner/repo` with the repository path.

```bash
gh run view <run-id> --repo owner/repo
```

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

### Check Git Tags

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Verify that the release tag v1.6.2 does not already exist in the remote repository. This ensures the tag is unique.

```bash
git ls-remote --tags origin v1.6.2
```

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

### Clone an Agent

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Clone an existing agent, including all its configuration. Use `--json` for structured output.

```bash
swarmclaw agents clone <agentId> --json
```

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

### List All Agents

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

List all agents in your SwarmClaw instance. Use `--json` for structured output.

```bash
swarmclaw agents list --json
```

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

### Run Codex in a Scratch Directory

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/coding-agent/SKILL.md

When Codex refuses to run outside a git directory, initialize a git repository in a temporary directory for scratch work.

```bash
SCRATCH=$(mktemp -d) && cd $SCRATCH && git init && codex exec "Your prompt"
```

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

### List Recent Workflow Runs

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

List the most recent workflow runs for a repository, limited to 10. Replace `owner/repo` with the repository path.

```bash
gh run list --repo owner/repo --limit 10
```

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

### Run Schedule

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Immediately triggers a specific schedule to execute its task. Useful for testing or on-demand execution.

```bash
swarmclaw schedules run <scheduleId> --json
```

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

### Run Knowledge Hygiene Maintenance

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Executes the knowledge hygiene maintenance process, automatically syncing stale sources and archiving duplicates. This command helps maintain the quality of knowledge sources.

```bash
swarmclaw knowledge hygiene-run --json
```

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

### List Memories

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Lists all stored memories, which can be searched by query parameters like content or agent ID. Essential for retrieving contextual information.

```bash
swarmclaw memory list --json
```

```bash
swarmclaw memory list --json --query q=pricing --query agentId=<id>
```

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

### Run CI Validation Checks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

Execute these commands to ensure all CI validation checks pass before committing a release. This includes linting, production build, and various test suites.

```bash
npm run lint:baseline
```

```bash
NODE_ENV=production npm run build:ci
```

```bash
npm run test:cli
```

```bash
npm run test:openclaw
```

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

### System Resource Usage

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Displays information about the resource usage of the SwarmClaw system. Useful for monitoring performance and capacity.

```bash
swarmclaw system usage --json
```

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

### Generate Images with DALL-E 3

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/openai-image-gen/SKILL.md

Utilize DALL-E 3 for image generation, supporting HD quality and specific styles like 'vivid' or 'natural'. Note that the count is automatically limited to 1 for DALL-E 3.

```bash
python3 {baseDir}/scripts/gen.py --model dall-e-3 --quality hd --size 1792x1024 --style vivid
```

```bash
python3 {baseDir}/scripts/gen.py --model dall-e-3 --style natural --prompt "serene mountain landscape"
```

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

### Fix Lint Blockers

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

Before preparing a release, run ESLint on changed files to catch issues. Update the lint baseline after fixing any identified problems.

```bash
npx eslint <changed-files>
```

```bash
npm run lint:baseline:update
```

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

### Generate PR Review Summary

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Generate a summary for a pull request including title, body, author, and file changes, followed by its CI checks. Set PR and REPO variables.

```bash
PR=55 REPO=owner/repo
echo "## PR #$PR Summary"
gh pr view $PR --repo $REPO --json title,body,author,additions,deletions,changedFiles \
  --jq "**\(.title)** by @\(.author.login)\n\n\(.body)\n\n+\(.additions) -\(.deletions) across \(.changedFiles) files"
gh pr checks $PR --repo $REPO
```

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

### List Mergeable Pull Requests

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

List pull requests, outputting number, title, state, and mergeability, then filter to show only mergeable ones. Replace `owner/repo` with the repository path.

```bash
gh pr list --json number,title,state,mergeable --jq '.[] | select(.mergeable == "MERGEABLE")'
```

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

### Take a Screenshot

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/browser.md

Capture a screenshot of the current browser view. This is useful for visual verification or debugging.

```json
{ "action": "screenshot" }
```

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

### Query GitHub API for Pull Request Details

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

Make an API query to retrieve specific fields (title, state, author login) of a pull request. Replace `owner/repo` and `55` accordingly.

```bash
gh api repos/owner/repo/pulls/55 --jq '.title, .state, .user.login'
```

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

### Run OpenAI Image Gen Script

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/openai-image-gen/SKILL.md

Execute the main image generation script. Ensure a higher timeout is set for longer generation tasks.

```bash
python3 {baseDir}/scripts/gen.py
```

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

### Create Memory Entry

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Creates a new memory entry for an agent, storing content and type. This is key for long-term context persistence.

```bash
swarmclaw memory create --data '{"agentId":"<id>","content":"Client prefers formal tone","type":"preference"}' --json
```

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

### Run Codex One-Shot (Auto-Approve)

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/coding-agent/SKILL.md

Execute Codex for quick, one-shot tasks where changes are automatically approved. Codex requires a git repository and PTY mode.

```bash
cd /path/to/project && codex exec --full-auto 'Build a dark mode toggle'
```

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

### Specify Aspect Ratio

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/nano-banana-pro/SKILL.md

Control the aspect ratio of the generated image using the `--aspect-ratio` flag. Supported ratios include 9:16, 16:9, and others.

```bash
uv run {baseDir}/scripts/generate_image.py --prompt "portrait photo" --filename "output.png" --aspect-ratio 9:16
```

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

### Bulk Update Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Perform a bulk update on multiple tasks by specifying their IDs and the changes to apply in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw tasks bulk --data '{"ids":["id1","id2"],"changes":{"status":"todo"}}' --json
```

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

### Projects API

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Access and retrieve information about projects.

```APIDOC
## projects.list

### Description
Lists all available projects.

### Method
Not applicable (action-based API)

### Endpoint
Not applicable (action-based API)

### Parameters
#### Action Parameters
- **action** (string) - Required - Must be `projects.list`

### Request Example
```json
{ "action": "projects.list" }
```

## projects.get

### Description
Retrieves details for a specific project.

### Method
Not applicable (action-based API)

### Endpoint
Not applicable (action-based API)

### Parameters
#### Action Parameters
- **action** (string) - Required - Must be `projects.get`
- **id** (string) - Required - The ID of the project to retrieve.
```

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

### Click an Element by Selector

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/browser.md

Click an element on the page using its CSS selector. Ensure the selector accurately targets the desired element.

```json
{ "action": "click", "selector": "button.submit" }
```

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

### Ask Human for Input

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Block agent execution and wait for a human response to a question. Provide context for clarity. Use sparingly when human input is essential for proceeding.

```json
{
  "action": "communicate.ask_human",
  "question": "Should I proceed with the database migration?",
  "context": "This will add 3 new columns to the users table and backfill existing rows."
}
```

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

### Test CLI Utilities

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Run specific tests for CLI utility functions using Node.js with tsx import. This targets regression checks.

```bash
node --import tsx --test src/lib/providers/cli-utils.test.ts
```

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

### Execute Bash Script (JSON Input)

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

This snippet shows how to execute a bash script using a JSON payload, often used for tool calls.

```json
{ "code": "curl -s https://api.example.com/data | jq '.results[]'" }
```

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

### View Pull Request Details

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

View detailed information about a specific pull request. Replace `55` with the PR number and `owner/repo` with the repository path.

```bash
gh pr view 55 --repo owner/repo
```

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

### Hosted MCP Surface

Source: https://github.com/swarmclawai/swarmclaw/blob/main/SWARMDOCK.md

The single hosted MCP endpoint that wraps marketplace tools for MCP-capable clients.

```APIDOC
## POST https://swarmdock-api.onrender.com/mcp

### Description
This is the single hosted MCP endpoint that wraps marketplace tools. It supports operations like registering, listing tasks, bidding, checking status, and updating profiles. Clients interact with this endpoint using JSON-RPC exchanges.

### Method
POST

### Endpoint
`https://swarmdock-api.onrender.com/mcp`

### Authentication
`Authorization: Bearer <agent Ed25519 secret, base64>`

### Transport
Stateless `StreamableHTTP` — each POST is a complete JSON-RPC exchange.
```

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

### Filter Memory List by Agent and Query

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Filter memory entries by agent ID and a specific query string. Use `--json` for structured output.

```bash
swarmclaw memory list --json --query agentId=abc123 --query q=pricing
```

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

### Send Message and Stream Response

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Send a message to a chat session and stream the agent's response using Server-Sent Events (SSE). Use `--json` for structured output.

```bash
swarmclaw chats chat <chatId> --data '{"message":"Give me a status update"}' --json
```

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

### Storage: Normalization on Load

Source: https://github.com/swarmclawai/swarmclaw/blob/main/CLAUDE.md

Apply storage normalization on load to automatically migrate records and apply default values for new fields. This avoids scattered `undefined` checks.

```typescript
// In storage-normalization.ts:
export function normalizeMyRecord(record: any): MyRecordWithDefaults {
  return {
    ...record,
    newField: record.newField ?? "defaultValue", // Apply default for new fields
    anotherField: record.anotherField ?? 0,
  };
}

// When loading records:
// const records = await storage.loadCollection("myCollection");
// const normalizedRecords = records.map(normalizeMyRecord);

```

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

### Spawn a Subagent

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Create a new subagent to perform tasks in parallel. Choose the 'run' mode for independent execution or 'session' for a persistent session.

```json
{
  "action": "communicate.spawn",
  "agentId": "agent_coder',
  "message": "Implement the dark mode toggle component",
  "mode": "run"
}
```

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

### Schedule Recurring Tasks

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Set up a recurring task for an agent, specifying the cron schedule and message content. You need the agent's ID.

```bash
swarmclaw agents list --json
```

```bash
# Find the reporting agent's ID
```

```bash
swarmclaw schedules create --data '{"name":"Daily Morning Summary","agentId":"<reporting-agent-id>","cron":"0 9 * * *","message":"Generate and send the daily summary report covering key metrics, alerts, and action items from the last 24 hours"}' --json
```

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

### File Inspection in Sandbox

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/execute.md

These bash commands are used for file inspection within the sandbox environment. The first counts TypeScript files, and the second recursively searches for 'TODO' comments in TypeScript files.

```bash
find /workspace/src -name "*.ts" | wc -l
grep -r "TODO" /workspace/src --include="*.ts" -l
```

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

### Supersede Knowledge Source

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Marks a knowledge source as superseded by another, indicating a replacement. Requires the ID of the new source.

```bash
swarmclaw knowledge source-supersede <sourceId> --data '{"supersededBySourceId":"<newSourceId>"}' --json
```

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

### Restore Archived Knowledge Source

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Restores a previously archived knowledge source, making it active again. Use the source ID to specify the source.

```bash
swarmclaw knowledge source-restore <sourceId> --json
```

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

### System Version Check

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Checks the current SwarmClaw version and available updates. Important for staying current with the latest features and fixes.

```bash
swarmclaw system version --json
```

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

### Search Resources

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Performs a global search across all SwarmClaw resources, including agents, tasks, chats, and more. Use the query parameter to specify search terms.

```bash
swarmclaw search query --json --query q=pricing
```

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

### Check npm Package Version

Source: https://github.com/swarmclawai/swarmclaw/blob/main/docs/release/v1.6.2-codex-continuity-plan.md

Verify that the specified version of the SwarmClaw package is not yet published to npm. This should return empty if the version is not available.

```bash
npm view @swarmclawai/swarmclaw@1.6.2 version
```

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

### View Failed Workflow Run Logs

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/github/SKILL.md

View only the logs for failed steps of a specific workflow run. Replace `<run-id>` with the run ID and `owner/repo` with the repository path.

```bash
gh run view <run-id> --repo owner/repo --log-failed
```

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

### Bulk Update Agents

Source: https://github.com/swarmclawai/swarmclaw/blob/main/swarmclaw/SKILL.md

Perform a bulk update on multiple agents by specifying their IDs and the changes to apply in JSON format via the `--data` flag. Use `--json` for structured output.

```bash
swarmclaw agents bulk-update --data '{"ids":["id1","id2"],"changes":{"provider":"openai"}}' --json
```

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

### Delegate Work to Another Agent

Source: https://github.com/swarmclawai/swarmclaw/blob/main/skills/tools/platform.md

Route a task or message to a different agent. Provide a clear message with sufficient context for the target agent to work independently.

```json
{
  "action": "communicate.delegate",
  "agentId": "agent_research",
  "message": "Find the top 5 competitors in the AI coding assistant space and summarize their pricing."
}
```
Tokens:
Raw