메뉴
Add a local Ollama MCP server so the Codex worker can call local models for cheaper/faster summarization, translation, or general helper tasks.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Add iMessage as an optional channel skill for NanoClaw using a single-file local-macos-first implementation. Phase 1 keeps BlueBubbles as a config seam.
Add new capabilities or modify NanoClaw behavior. Use when user wants to add channels (iMessage, Telegram, Slack, email input), change triggers, add integrations, modify the router, or make any other customizations. This is an interactive skill that asks questions to understand what the user wants.
Run initial NanoClaw setup. Use when user wants to install dependencies, authenticate messaging channels, register their main channel, or start the background services. Triggers on "setup", "install", "configure nanoclaw", or first-time setup requests.
Add Gmail integration to NanoClaw. Can be configured as a tool (agent reads/sends emails when triggered from WhatsApp) or as a full channel (emails can trigger the agent, schedule tasks, and receive replies). Guides through GCP OAuth setup and implements the integration.
Add Slack as a channel. Can replace WhatsApp entirely or run alongside it. Uses Socket Mode (no public URL needed).
Add Telegram as a channel. Can replace WhatsApp entirely or run alongside it. Also configurable as a control-only channel (triggers actions) or passive channel (receives notifications only).
| name | add-ollama-tool |
| description | Add a local Ollama MCP server so the Codex worker can call local models for cheaper/faster summarization, translation, or general helper tasks. |
This skill adds a stdio MCP server that exposes local Ollama models to the local Codex worker.
Tools added:
ollama_list_models — lists installed Ollama modelsollama_generate — sends a prompt to a model and returns the responseNanoClaw now runs a local Codex worker, not a Docker container. Avoid Docker-only assumptions such as host.docker.internal, per-group agent-runner-src copies, or rebuilding container/build.sh.
The skill package may still contain old container-era modify/ files. If an apply operation conflicts, port the intended changes into the current local-worker files:
container/agent-runner/src/ollama-mcp-stdio.tscontainer/agent-runner/src/runtime/codex-runtime.tssrc/container-runner.ts only if host log surfacing is still neededRead .nanoclaw/state.yaml. If ollama is in applied_skills, skip to configuration and verification.
Verify Ollama is installed and running on the host:
ollama list
If Ollama is not installed, direct the user to https://ollama.com/download.
If no models are installed, suggest pulling one:
ollama pull gemma3:1b
ollama pull llama3.2
ollama pull qwen3-coder:30b
Initialize the skills system if needed:
npx tsx scripts/apply-skill.ts --init
Apply the skill:
npx tsx scripts/apply-skill.ts .claude/skills/add-ollama-tool
Expected outcome:
container/agent-runner/src/ollama-mcp-stdio.tsollama stdio MCP server in the Codex runtime's mcp_servers configOLLAMA_HOST.nanoclaw/state.yamlValidate:
npm run build
By default, local worker processes can use:
OLLAMA_HOST=http://127.0.0.1:11434
Add a different OLLAMA_HOST to .env only when Ollama is not listening on localhost.
Restart the service:
launchctl kickstart -k gui/$(id -u)/com.nanoclaw
# Linux: systemctl --user restart nanoclaw
Tell the user to ask from a registered chat:
use ollama to list available local models
Then monitor:
tail -f logs/nanoclaw.log | grep -i ollama
ls -lt groups/*/logs/worker-*.log | head
ollama listOLLAMA_HOST in .envcurl -s http://127.0.0.1:11434/api/tagscontainer/agent-runner/src/runtime/codex-runtime.ts includes the ollama MCP server entry.npm run build.groups/<group>/logs/worker-*.log for MCP startup errors.