Fan Manager

CLI | MCP | Agent

Version: 2.0.0

Documentation — Installation, deployment, and usage across the CLI and MCP interfaces, plus the integrated A2A agent server, are maintained in the official documentation.

Overview

Fan Manager controls the fan speed of Dell PowerEdge servers based on CPU temperature. It is a local tool: it reads temperatures via lm-sensors (sensors -j) and drives the server's BMC via ipmitool raw commands. It ships as a CLI, a Model Context Protocol (MCP) server, and an integrated A2A agent for the agent-utilities ecosystem.

Because fan control happens against the local host, no service URL or token is required — the connector degrades to a no-op/local config for authentication.

Key Features

• Temperature-Driven Curve: Logarithmic CPU-temperature-to-fan-speed scaling with configurable min/max bounds and intensity.
• Consolidated Action-Routed MCP Tools: Two togglable tool modules (temperature, fan-control) minimize token overhead in LLM contexts.
• Fail-Safe Defaults: On a temperature read error during automatic control, the fans default to maximum.
• Integrated Agent: Built-in Pydantic AI agent supporting the Agent Control Protocol (ACP) and Web UI (AG-UI).

CLI

The classic continuous service that polls temperature and adjusts the fans:

fan-manager --intensity 5 --cold 50 --warm 80 --slow 5 --fast 100 --poll-rate 24

Requires ipmitool and lm-sensors installed on the host, and privileges to issue raw IPMI commands to the BMC.

MCP

This server uses dynamic Action-Routed tools to optimize token overhead and maximize IDE compatibility.

Available MCP Tools

Auto-generated from the live MCP server — do not edit by hand.

Condensed action-routed tools (MCP_TOOL_MODE=condensed)

Verbose 1:1 API-mapped tools (MCP_TOOL_MODE=verbose or both)

9 action-routed tool(s) · 5 verbose 1:1 tool(s). Each is enabled unless its <DOMAIN>TOOL toggle is set false; MCP_TOOL_MODE selects the surface (intent default — the six verb-tools, granular set loaded on demand · condensed action-routed · verbose 1:1 · both). Auto-generated — do not edit.

Dynamic Tool Selection & Visibility

This MCP server supports dynamic toolset selection and visibility filtering at runtime, so you can restrict the exposed tools and avoid blowing up the LLM's context window. Configure filtering via:

• CLI Arguments: --tools / --toolsets (and --disabled-tools / --disabled-toolsets).
• Environment Variables: MCP_ENABLED_TOOLS / MCP_DISABLED_TOOLS, MCP_ENABLED_TAGS / MCP_DISABLED_TAGS.
• HTTP Headers: x-mcp-enabled-tools / x-mcp-disabled-tags, etc.
• Query Parameters: ?tools=tool1,tool2 or ?tags=fan-control.

MCP Configuration Examples

Install the connector-focused [mcp] extra. Examples use fan-manager[mcp] to add FastMCP / FastAPI through agent-utilities[mcp]; the required Agent Utilities core still carries epistemic-graph[full]. The [agent-runtime] extra additionally enables model orchestration.

stdio Transport (local IDEs — Cursor, Claude Desktop, VS Code)

{
  "mcpServers": {
    "fan-manager-mcp": {
      "command": "uvx",
      "args": [
        "--from",
        "fan-manager[mcp]",
        "fan-manager-mcp"
      ],
      "env": {
        "MCP_TOOL_MODE": "intent",
        "ENABLE_DELEGATION": "False",
        "FAN_CONTROLTOOL": "True",
        "FAN_MANAGER_HOSTS": "r510,r710,r820,rw710",
        "FAN_MANAGER_KG_AGGREGATE_S": "3600",
        "FAN_MANAGER_KG_INGEST": "true",
        "FAN_MANAGER_POLICY_FILE": "/policy/fan-policy.json",
        "FAN_MANAGER_POLICY_REFRESH": "20",
        "IPMITOOL": "True",
        "IPMITOOL_PATH": "ipmitool",
        "KGTOOL": "True",
        "SENSORS_PATH": "sensors",
        "TEMPERATURETOOL": "True"
      }
    }
  }
}

Runtime references require an alias-aware launcher such as GraphOS. Other launchers must omit those entries and inject the resolved values through their own runtime secret boundary.

Streamable-HTTP Transport (networked / production)

{
  "mcpServers": {
    "fan-manager-mcp": {
      "command": "uvx",
      "args": [
        "--from",
        "fan-manager[mcp]",
        "fan-manager-mcp",
        "--transport",
        "streamable-http",
        "--port",
        "8000"
      ],
      "env": {
        "TRANSPORT": "streamable-http",
        "HOST": "127.0.0.1",
        "PORT": "8000",
        "MCP_TOOL_MODE": "intent",
        "ENABLE_DELEGATION": "False",
        "FAN_CONTROLTOOL": "True",
        "FAN_MANAGER_HOSTS": "r510,r710,r820,rw710",
        "FAN_MANAGER_KG_AGGREGATE_S": "3600",
        "FAN_MANAGER_KG_INGEST": "true",
        "FAN_MANAGER_POLICY_FILE": "/policy/fan-policy.json",
        "FAN_MANAGER_POLICY_REFRESH": "20",
        "IPMITOOL": "True",
        "IPMITOOL_PATH": "ipmitool",
        "KGTOOL": "True",
        "SENSORS_PATH": "sensors",
        "TEMPERATURETOOL": "True"
      }
    }
  }
}

Alternatively, connect to a pre-deployed Streamable-HTTP instance by url:

{
  "mcpServers": {
    "fan-manager-mcp": {
      "url": "http://localhost:8000/fan-manager-mcp/mcp"
    }
  }
}

Run a reviewed container image as a least-privilege stdio child (no listener or published port):

docker run -i --rm \
  --read-only \
  --cap-drop=ALL \
  --security-opt=no-new-privileges \
  --pids-limit=256 \
  --tmpfs /tmp:rw,noexec,nosuid,nodev,size=64m \
  -e TRANSPORT=stdio \
  -e MCP_TOOL_MODE=intent \
  -e ENABLE_DELEGATION=False \
  -e FAN_CONTROLTOOL=True \
  -e FAN_MANAGER_HOSTS=r510,r710,r820,rw710 \
  -e FAN_MANAGER_KG_AGGREGATE_S=3600 \
  -e FAN_MANAGER_KG_INGEST=true \
  -e FAN_MANAGER_POLICY_FILE=/policy/fan-policy.json \
  -e FAN_MANAGER_POLICY_REFRESH=20 \
  -e IPMITOOL=True \
  -e IPMITOOL_PATH=ipmitool \
  -e KGTOOL=True \
  -e SENSORS_PATH=sensors \
  -e TEMPERATURETOOL=True \
  registry.example.invalid/fan-manager@sha256:<digest> fan-manager-mcp

For containerized network HTTP, supply an authenticated TLS ingress (or direct server TLS), exact MCP_ALLOWED_HOSTS, and an exact trusted-proxy CIDR policy through the operator-owned deployment profile. The generator does not emit an unauthenticated non-loopback listener.

Auto-generated from the code-read env surface (MCP_TOOL_MODE + package vars) — do not edit.

Additional Deployment Options

fan-manager can run as a local stdio process or container, or behind a remote network boundary. The Deployment guide carries the detailed transport contract.

• Local container — launch a reviewed immutable image as a least-privilege stdio child with no listener or published port.
• Remote URL — connect through an operator-supplied authenticated HTTPS ingress. Keep its URL, outbound identity references, trust profile, and exact MCP_ALLOWED_HOSTS in AgentConfig.

Agent

This repository features a fully integrated Pydantic AI Graph Agent. It communicates over the Agent Control Protocol (ACP) and interacts with the Agent Web UI (AG-UI) and Terminal interface.

Running the Agent CLI

fan-manager-agent --provider openai --model-id gpt-4o

Docker Compose Orchestration

The docker/agent.compose.yml configures the Agent, Web UI, and Terminal Interface alongside the MCP server. See docs/deployment.md for the full Compose stack.

Environment Variables

Package environment variables

Inherited agent-utilities variables (apply to every connector)

28 package + 16 inherited variable(s). Auto-generated from .env.example + the shared agent-utilities set — do not edit.

Fan Manager reads the following environment variables (all optional — every one has a safe default). They can be set in the process environment, in a .env file (auto-loaded), or in the MCP client's env block. See .env.example for a ready-to-copy template.

Build-time only (not application config): UV_COMPILE_BYTECODE, NO_COLOR, and TERM are consumed by the build/runtime environment (Docker image build, terminal rendering) and are not read by the application code.

Security & Governance

Built directly upon the enterprise-ready agent-utilities core, standard security parameters are fully supported:

Access Control & Policy Enforcement

• Eunomia Policies: Fine-grained, policy-driven tool authorization (none, embedded, or remote).
• OIDC Token Delegation: Optional RFC 8693 token exchange (inert by default — Fan Manager is a local tool).

Runtime Security Grid

Installation

Pick the extra that matches what you want to run:

# Connector-focused MCP server (includes the shared graph engine)
uv pip install "fan-manager[mcp]"

# Agent runtime (adds model orchestration to the shared graph engine)
uv pip install "fan-manager[agent]"

# Everything (development)
uv pip install "fan-manager[all]"      # or: python -m pip install "fan-manager[all]"

Container images (:mcp vs :agent)

One multi-stage docker/Dockerfile builds two right-sized images, selected by --target:

docker build --target mcp   -t example/fan-manager:mcp    docker/   # connector-focused MCP server
docker build --target agent -t example/fan-manager:agent-local docker/   # agent runtime

docker/mcp.compose.yml runs the connector-focused :mcp server; docker/agent.compose.yml runs the agent (immutable agent digest) with a co-located :mcp sidecar.

Knowledge-graph database (epistemic-graph)

Both [mcp] and [agent] carry the epistemic-graph engine through the required Agent Utilities core dependency (epistemic-graph[full]). The [mcp] extra keeps the server connector-focused; [agent] additionally enables model orchestration. Local deployments can use the bundled engine. For production or shared state, run epistemic-graph as a dedicated database service and configure the runtime to use it. Deployment recipes (single-node + Raft HA), connection configuration, and architecture diagrams are documented in the epistemic-graph deployment guide.

Documentation

The complete documentation is published as the official documentation site.

Contribute

Contributions are welcome! Please ensure code quality by executing local checks before submitting pull requests:

• Format code using ruff format .
• Lint code using ruff check .
• Validate type-safety with mypy .
• Execute test suites using pytest

Deploy with agent-utilities-deployment

Provision this package with the consolidated agent-utilities-deployment workflow. It selects an installed-package, editable-source, or immutable-container path; records only runtime secret and TLS-profile references in AgentConfig; and runs doctor, registration, policy, observability, and rollback gates. Ask your agent to "deploy fan-manager with agent-utilities-deployment".

The repository embeds no deployment profile, credential value, certificate path, or environment-specific endpoint. Supply those at runtime through AgentConfig and the configured secret provider.

Governed capability contract

This package ships a compact canonical skill surface with specialist procedures kept as referenced workflows. The current MCP tools, skill metadata, connector_manifest.yml, ontology, mappings, shapes, fixtures, migrations, tool-schema fingerprints, and certification metadata form one versioned capability contract. Validate them together; do not rely on stale tool names or historical per-task skill wrappers.

Runtime endpoints, credentials, certificate trust, tenant identity, retention, and observability policy are deployment inputs and are never packaged values. See Configuration, trust, and privacy before enabling a network transport, connector ingestion, GraphOS delegation, or trace export.