- Host-side workspace skills (
.claude/skills/) — invoked as slash commands in Claude Code from your NanoClaw checkout. They install channels, providers, and tools, or operate and maintain your install. 45 skills total: 17 channel installs, 3 provider installs, 10 tool installs, and 15 operational skills. - Container skills (
container/skills/) — 8 skills mounted into every agent container at/app/skills. The agent loads them at runtime; you don’t invoke them yourself.
Channel installs
Each skill wires up one messaging channel. See Channels for the full setup guides.| Skill | What it does |
|---|---|
/add-deltachat | Adds the Delta Chat channel via @deltachat/stdio-rpc-server — native adapter, email-based messaging with end-to-end encryption |
/add-discord | Adds the Discord bot channel via Chat SDK |
/add-emacs | Adds Emacs as a channel — interactive chat buffer and org-mode integration over a local HTTP bridge, no bot token needed |
/add-gchat | Adds the Google Chat channel via Chat SDK |
/add-github | Adds the GitHub channel via Chat SDK — PR and issue comment threads as conversations |
/add-imessage | Adds the iMessage channel via Chat SDK — local (macOS) or remote (Photon API) mode |
/add-linear | Adds the Linear channel via Chat SDK — issue comment threads as conversations |
/add-matrix | Adds the Matrix channel via Chat SDK — works with any homeserver |
/add-resend | Adds the Resend email channel via Chat SDK |
/add-signal | Adds the Signal channel via signal-cli TCP daemon — native adapter, no Chat SDK bridge |
/add-slack | Adds the Slack channel via Chat SDK |
/add-teams | Adds the Microsoft Teams channel via Chat SDK |
/add-telegram | Adds the Telegram channel via Chat SDK |
/add-webex | Adds the Webex channel via Chat SDK |
/add-wechat | Adds the WeChat personal channel via Tencent’s official iLink Bot API — long-polling and QR scan, no webhook |
/add-whatsapp | Adds the WhatsApp channel via native Baileys adapter — QR code or pairing code authentication |
/add-whatsapp-cloud | Adds the WhatsApp Business Cloud API channel via Chat SDK — official Meta API |
Provider installs
Each skill swaps the agent runtime for a different provider. See Providers.| Skill | What it does |
|---|---|
/add-codex | Uses Codex (CLI + AppServer) as the full agent provider in place of the Claude Agent SDK — ChatGPT subscription or OPENAI_API_KEY, per-group via agent_provider |
/add-ollama-provider | Routes an agent group to a local Ollama model instead of the Anthropic API — env var overrides and a model setting, no provider code changes |
/add-opencode | Uses OpenCode as an agent provider (AGENT_PROVIDER=opencode) — OpenRouter, OpenAI, Google, DeepSeek, and more via OpenCode config |
Tool installs
Each skill adds a capability or MCP tool to your agents. See Tools.| Skill | What it does |
|---|---|
/add-atomic-chat-tool | Adds the Atomic Chat MCP server so agents can call local models served by the Atomic Chat desktop app via its OpenAI-compatible API |
/add-dashboard | Adds a monitoring dashboard — installs @nanoco/nanoclaw-dashboard and a pusher that sends periodic JSON snapshots |
/add-gcal-tool | Adds Google Calendar as an MCP tool (list calendars, list/search/create events, free/busy) using OneCLI-managed OAuth — no raw credentials reach the container |
/add-gmail-tool | Adds Gmail as an MCP tool (read, search, send, label, draft) using OneCLI-managed OAuth — tokens injected at request time |
/add-karpathy-llm-wiki | Adds a persistent wiki knowledge base to a group, based on Karpathy’s LLM Wiki pattern |
/add-macos-statusbar | Adds a macOS menu bar status indicator with Start, Stop, and Restart controls (macOS only) |
/add-mnemon | Adds persistent graph-based memory via mnemon — agents recall past context before responding and remember insights after each turn |
/add-ollama-tool | Adds the Ollama MCP server so agents can call local models and optionally manage the Ollama model library |
/add-rtk | Installs the rtk token-compression proxy into agent containers — routes Bash tool calls through rtk for 60–90% token savings on dev commands |
/add-vercel | Adds Vercel deployment capability — installs the Vercel CLI in agent containers with OneCLI credential injection for api.vercel.com |
Operational skills
Setup, maintenance, migration, and day-to-day administration of your install.| Skill | What it does |
|---|---|
/customize | Adds new capabilities or modifies NanoClaw behavior — interactive skill that asks questions to understand what you want |
/debug | Debugs container agent issues — logs, session DBs, mounts, authentication problems, and common failures |
/get-qodo-rules | Loads org- and repo-level coding rules from Qodo before code tasks begin, so generation follows team standards |
/init-first-agent | Walks you through creating the first agent for a DM channel — resolves your channel identity, wires the DM group, and triggers a welcome DM |
/init-onecli | Installs and initializes OneCLI Agent Vault, migrating existing .env credentials to the vault |
/manage-channels | Wires channels to agent groups, manages isolation levels, and adds new channel groups |
/manage-mounts | Configures which host directories agent containers can access — view, add, or remove mount allowlist entries |
/migrate-from-openclaw | Migrates from OpenClaw to NanoClaw v2 — detects an existing install, extracts identity, credentials, and tasks, then guides interactive migration |
/migrate-from-v1 | Finishes migrating a NanoClaw v1 install into v2 after migrate-v2.sh — seeds the owner, reconciles container configs, helps port custom v1 code |
/migrate-nanoclaw | Extracts your customizations from a fork, generates a replayable migration guide, and upgrades to upstream by reapplying them on a clean base |
/qodo-pr-resolver | Reviews and resolves PR issues with Qodo — AI-powered code review fixes for GitHub, GitLab, Bitbucket, and Azure DevOps |
/setup | Runs initial NanoClaw setup — install, configure, first-time setup |
/update-nanoclaw | Brings upstream updates into a customized install with preview, selective cherry-pick, and low token usage |
/update-skills | Re-applies your installed skills to pull their latest code from upstream |
/use-native-credential-proxy | Opts out of the OneCLI gateway — reads Anthropic credentials from .env and injects them into the container’s API requests |
Container skills
Mounted into every agent container at/app/skills. The agent picks them up automatically based on context — channel, task, or trigger.
| Skill | What it does |
|---|---|
agent-browser | Browses the web for any task — research, web apps, forms, screenshots, data extraction, page testing |
frontend-engineer | Enforces a build-test-verify workflow for web projects — never done until built, tested, responsive, accessible, and visually verified |
onecli-gateway | Transparent HTTPS proxy that injects stored credentials into outbound calls — the agent makes plain HTTP requests and the gateway handles auth |
self-customize | Lets the agent customize itself — add capabilities, install packages, add MCP servers, edit code or CLAUDE.md |
slack-formatting | Formats messages for Slack using mrkdwn syntax when responding to Slack channels |
vercel-cli | Deploys apps to Vercel and manages Vercel projects, domains, and environment variables |
welcome | Introduces the agent to a newly connected channel — triggered automatically when a channel is first wired |
whatsapp-formatting | Formats messages for WhatsApp, including mentions that render as real WhatsApp tags |
Related pages
- Extending NanoClaw — how skills work and how to write your own
- Channels overview — channel setup guides
- Providers — agent provider concepts
- Tools — adding MCP tools to agents