Skills and Plugins

What Are Skills?

Skills are pre-built capabilities that extend what your OpenClaw instance can do. A skill teaches your instance how to perform a specific task — like searching the web, reading documents, generating images, or interacting with external APIs.

OpenClaw ships with a set of built-in skills. When you enable certain add-ons or configure messenger channels, additional skills become available automatically.

What Are Plugins?

Plugins are the runtime components that connect your instance to external services. When you add a Telegram bot token, the Telegram plugin handles receiving and sending messages. When you configure an LLM API key, the LLM plugin routes requests to your chosen provider.

The distinction matters mostly under the hood. From a user perspective: skills define what the instance can do, plugins handle the connections that make it possible.

How Skills and Plugins Get Activated

You do not install skills or plugins manually. ClawHosters handles activation automatically:

1. You configure a feature in your dashboard (add Telegram, set up LLM, enable voice)
2. ClawHosters updates your instance's openclaw.json configuration
3. The doctor --fix command runs on your instance
4. The doctor scans the config and enables the corresponding skills and plugins
5. Your instance picks up the changes within seconds

If you change your configuration later (switch LLM provider, add Discord), the same process runs again and adjusts which skills and plugins are active.

Built-in Skills

Every OpenClaw instance includes these skills by default, regardless of configuration:

Skill	What It Does
Conversation	Handles back-and-forth chat with users
Memory	Remembers context across conversations
Commands	Processes slash commands from users
System	Handles instance health checks and diagnostics

Add-on Skills

These skills activate when you subscribe to the corresponding add-on:

Add-on	Skills Enabled
LLM	AI response generation, reasoning, code execution
Voice	Speech-to-text transcription, text-to-speech output
Embeddings	Document search, knowledge base queries, RAG
Backup	Snapshot creation, restore from backup

Without the LLM add-on, the instance runs in gateway-only mode — it can receive messages but cannot generate AI responses.

Channel Plugins

Each messenger channel has its own plugin:

Channel	Plugin	Activated When
Telegram	Telegram polling plugin	Bot token configured
Discord	Discord event listener	Bot token configured
Slack	Slack socket mode plugin	Bot + app tokens configured
WhatsApp	WhatsApp bridge plugin	WhatsApp connected

Channel plugins start automatically when their tokens are present in openclaw.json. They stop when tokens are removed.

Installing Custom Skills via SSH

If you SSH into your instance, you can install additional skills beyond what ClawHosters configures. Custom skills are placed in the instance's skills directory inside the container.

Keep in mind:

• Custom skills persist across container restarts (stored in the data volume)
• A rebuild wipes custom skills — you need to reinstall them after
• ClawHosters does not manage or support custom skills
• Broken custom skills can affect instance stability

For most users, the built-in and add-on skills cover standard use cases.

Troubleshooting

Skills Not Working After Config Change

If you changed a setting but the skill does not seem active:

1. Check the instance status on your dashboard — it should show "Running"
2. Wait 30 seconds for the config reload to complete
3. If still not working, try the "Redeploy Config" button in your dashboard
4. Check that your add-on subscription is active (not expired or paused)

Plugin Connection Errors

If a messenger plugin fails to connect:

• Verify the bot token is correct (copy-paste errors are common)
• For Telegram, make sure the bot is not already running elsewhere
• For Discord, check that the bot has the required permissions in your server
• For Slack, both the bot token and app-level token must be valid

Related Documentation

• OpenClaw Configuration Reference — Full config file reference
• LLM Add-on — Setting up AI responses
• Connecting Telegram — Telegram bot setup
• SSH Access — Connecting to your instance