add-ollama-tool

This Skill integrates local Ollama models into your containerized agent by registering an MCP server, enabling local model inference, model library management, and secure host-to-container environment variable forwarding.

Yes, you can manage your Ollama model library directly through the agent interface to optionally pull, delete, and inspect local models, ensuring your containerized agent has access to required inference resources.

Yes, using local Ollama model inference requires a running Ollama daemon on the host machine and proper registration of the stdio-based MCP server within the agent runner to establish communication.

Use local Ollama models for code generation to keep sensitive code tasks entirely on your local machine, avoiding cloud API costs and privacy concerns while leveraging specialized models like qwen3-coder.

This approach is limited to environments with a running Ollama daemon on the host and requires proper registration of the stdio-based MCP server within the agent runner to maintain secure, keyless local AI operations.