mcp-docker-deployment
Set up Docker deployment for Python MCP servers (FastMCP or low-level mcp.server.Server SDK) with SSE/streamable-http transport, automated versioning, and container registry publishing. Use when dockerizing an MCP server, containerizing for remote access, deploying an MCP server behind nginx, or setting up a production MCP server with Docker. Covers Dockerfile, build scripts, docker-compose, and nginx reverse proxy for SSE streaming.
What this skill does
# MCP Docker Deployment
Containerize Python MCP servers (FastMCP or low-level SDK) for remote deployment with SSE or streamable-http transport, nginx reverse proxy with HTTPS, and GHCR publishing.
## Step 1: Gather Project Information
Ask the user:
1. **"What is your MCP server entry point file?"** (e.g., `my_mcp_server.py`)
2. **"Does your server use FastMCP (`mcp.server.fastmcp.FastMCP`) or the low-level SDK (`mcp.server.Server`)?"**
3. **"Which transport? `sse` or `streamable-http`?"** (They are NOT interchangeable — different endpoints, different SDK classes)
4. **"What Python files/directories need to be copied into the container?"** (e.g., `tools/`, `models/`, `formatting.py`)
5. **"What container registry URL?"** (e.g., `ghcr.io/{org}/{project}`)
6. **"Does the MCP server need any environment variables?"** (list them for docker-compose and example.env)
7. **"Will this be behind an nginx reverse proxy with HTTPS?"** (yes/no - if yes, include nginx config)
8. **"Does it have private pip dependencies?"** (yes/no - if yes, needs CR_PAT build arg)
## Step 2: Create Dockerfile
```dockerfile
FROM python:3.13-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application code - adjust to match project structure
COPY {entry_point} .
# COPY additional files/directories as needed
# Non-root user
RUN useradd --create-home appuser && chown -R appuser:appuser /app
USER appuser
EXPOSE 8000
# Transport for Docker, bind to all interfaces
ENV MCP_TRANSPORT=sse
ENV MCP_HOST=0.0.0.0
ENV MCP_PORT=8000
CMD ["python", "{entry_point}"]
```
> **FastMCP note**: If the server uses FastMCP, also set `FASTMCP_HOST` and `FASTMCP_PORT` in the Dockerfile since FastMCP reads those specific env vars:
> ```dockerfile
> ENV FASTMCP_HOST=0.0.0.0
> ENV FASTMCP_PORT=8000
> ```
If private pip dependencies, add before `pip install`:
```dockerfile
ARG CR_PAT
RUN apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/* \
&& git config --global url."https://${CR_PAT}@github.com/".insteadOf "https://github.com/" \
&& pip install --no-cache-dir -r requirements.txt \
&& git config --global --unset url."https://${CR_PAT}@github.com/".insteadOf
```
## Step 3: Configure Transport in MCP Server
### Path A: FastMCP (`mcp.server.fastmcp.FastMCP`)
FastMCP's constructor defaults override env vars, so host/port **must** be passed explicitly:
```python
import os
from mcp.server.fastmcp import FastMCP
mcp = FastMCP(
"my_mcp_server",
host=os.getenv("FASTMCP_HOST", "127.0.0.1"),
port=int(os.getenv("FASTMCP_PORT", "8000")),
stateless_http=True,
)
# ... register tools ...
if __name__ == "__main__":
transport = os.getenv("MCP_TRANSPORT", "stdio")
mcp.run(transport=transport)
```
**CRITICAL**: Without explicit `host`/`port` args, the container binds to `127.0.0.1` and is unreachable despite `FASTMCP_HOST=0.0.0.0` being set. This is because FastMCP's pydantic-settings defaults take precedence over env vars when constructor args are provided.
**CRITICAL**: `stateless_http=True` is required when running behind a reverse proxy (nginx). Without it, the server tracks sessions via `Mcp-Session-Id` headers. If the proxy drops that header or the SSE connection breaks, clients get `"Session not found"` errors. Stateless mode makes each request independent, which is the correct mode for containerized deployments behind a proxy.
Supported transports:
- `stdio` - Local development (Claude Code local MCP servers)
- `sse` - Server-Sent Events, works behind nginx reverse proxy
- `streamable-http` - Newer HTTP transport
### Path B: Low-level SDK (`mcp.server.Server`)
The low-level SDK requires manual transport wiring with Starlette and uvicorn. Each transport type uses different SDK classes and different endpoints.
#### Required imports
```python
import asyncio
import contextlib
import os
from collections.abc import AsyncIterator
from typing import Any
import uvicorn
from mcp.server import Server
from mcp.server.sse import SseServerTransport
from mcp.server.stdio import stdio_server
from mcp.server.streamable_http_manager import StreamableHTTPSessionManager
from starlette.applications import Starlette
from starlette.responses import Response
from starlette.routing import Mount, Route
```
#### Server instance
```python
app = Server("my_mcp_server")
# ... register tools with @app.list_tools(), @app.call_tool(), etc. ...
```
#### stdio transport (local development)
```python
async def main_stdio() -> None:
"""Run the MCP server over stdio transport (local development)."""
async with stdio_server() as (read_stream, write_stream):
await app.run(
read_stream,
write_stream,
app.create_initialization_options()
)
```
#### SSE transport
```python
def main_sse() -> None:
"""Run the MCP server over SSE transport (Docker/remote deployment)."""
host: str = os.environ.get("MCP_HOST", "0.0.0.0")
port: int = int(os.environ.get("MCP_PORT", "8000"))
sse_transport = SseServerTransport("/messages/")
async def handle_sse(request: Any) -> Response:
async with sse_transport.connect_sse(
request.scope, request.receive, request._send
) as streams:
await app.run(
streams[0],
streams[1],
app.create_initialization_options(),
)
return Response()
starlette_app = Starlette(
routes=[
Route("/sse", endpoint=handle_sse),
Mount("/messages/", app=sse_transport.handle_post_message),
],
)
uvicorn.run(starlette_app, host=host, port=port)
```
#### streamable-http transport
```python
def main_streamable_http() -> None:
"""Run the MCP server over streamable-http transport."""
host: str = os.environ.get("MCP_HOST", "0.0.0.0")
port: int = int(os.environ.get("MCP_PORT", "8000"))
session_manager = StreamableHTTPSessionManager(
app=app,
json_response=False,
stateless=True,
)
@contextlib.asynccontextmanager
async def lifespan(starlette_app: Starlette) -> AsyncIterator[None]:
async with session_manager.run():
yield
starlette_app = Starlette(
routes=[
Mount("/mcp", app=session_manager.handle_request),
],
lifespan=lifespan,
)
uvicorn.run(starlette_app, host=host, port=port)
```
#### Transport selection
```python
if __name__ == "__main__":
transport: str = os.environ.get("MCP_TRANSPORT", "stdio")
if transport == "stdio":
asyncio.run(main_stdio())
elif transport == "sse":
main_sse()
elif transport == "streamable-http":
main_streamable_http()
else:
raise SystemExit(f"Unknown MCP_TRANSPORT: {transport} (expected: stdio, sse, streamable-http)")
```
## Step 4: Create build-publish.sh
```bash
#!/bin/sh
# Build and publish MCP Docker image
# Usage: ./build-publish.sh [--no-cache]
REGISTRY="{registry_url}"
NO_CACHE=""
if [ "$1" = "--no-cache" ]; then
NO_CACHE="--no-cache"
fi
if [ ! -f VERSION ]; then
echo "1" > VERSION
fi
CURRENT_VERSION=$(cat VERSION)
case "$CURRENT_VERSION" in
''|*[!0-9]*)
echo "ERROR: VERSION file contains non-numeric value: $CURRENT_VERSION"
exit 1
;;
esac
NEXT_VERSION=$((CURRENT_VERSION + 1))
echo "Building ${REGISTRY}:${NEXT_VERSION}..."
docker build \
--platform linux/amd64 \
$NO_CACHE \
-t "${REGISTRY}:${NEXT_VERSION}" \
.
if [ $? -ne 0 ]; then
echo "ERROR: Docker build failed"
exit 1
fi
docker tag "${REGISTRY}:${NEXT_VERSION}" "${REGISTRY}:latest"
echo "Pushing ${REGISTRY}:${NEXT_VERSION}..."
docker push "${REGISTRY}:${NEXT_VERSION}"
if [ $? -ne 0 ]; then
echo "ERROR: Push failed"
exit 1
fi
echo "Pushing ${REGISTRY}:latest..."
docker push "${REGISTRY}:latest"
if [ $? -ne 0 ]; then
echo "ERROR: Push failed"
exit 1
fiRelated in Backend & APIs
jfrog
IncludedInteract with the JFrog Platform via the JFrog CLI and REST/GraphQL APIs. Use this skill when the user wants to manage Artifactory repositories, upload or download artifacts, manage builds, configure permissions, manage users and groups, work with access tokens, configure JFrog CLI servers, search artifacts, manage properties, set up replication, manage JFrog Projects, run security audits or scans, look up CVE details, query exposures scan results from JFrog Advanced Security, manage release bundles and lifecycle operations, aggregate or export platform data, or perform any JFrog Platform administration task. Also use when the user mentions jf, jfrog, artifactory, xray, distribution, evidence, apptrust, onemodel, graphql, workers, mission control, curation, advanced security, exposures, or any JFrog product name.
cupynumeric-migration-readiness
IncludedPre-migration readiness assessor for porting NumPy to cuPyNumeric. Use BEFORE substantial porting work begins when the user asks whether code will scale on GPU, whether they should migrate to cuPyNumeric, which NumPy patterns transfer cleanly, what must be refactored before porting, or mentions pre-port assessment, scaling analysis, or refactor planning. Inspect the user's source code, look up NumPy usage, cross-reference the cuPyNumeric API support manifest, and distinguish distributed-scaling-friendly patterns from blockers such as unsupported APIs, scalar synchronization, host round-trips, Python/object-heavy control flow, shape/data-dependent branching, and in-place mutation hazards. Produce a verdict of READY, LIGHT REFACTOR, SIGNIFICANT REFACTOR, or NOT RECOMMENDED, with concrete refactor pointers.
alibabacloud-data-agent-skill
IncludedInvoke Alibaba Cloud Apsara Data Agent for Analytics via CLI to perform natural language-driven data analysis on enterprise databases. Data Agent for Analytics is an intelligent data analysis agent developed by Alibaba Cloud Database team for enterprise users. It automatically completes requirement analysis, data understanding, analysis insights, and report generation based on natural language descriptions. This tool supports: discovering data resources (instances/databases/tables) managed in DMS, initiating query or deep analysis sessions, real-time progress tracking, and retrieving analysis conclusions and generated reports. Use this Skill when users need to query databases, analyze data trends, generate data reports, ask questions in natural language, or mention "Data Agent", "data analysis", "database query", "SQL analysis", "data insights".
token-optimizer
IncludedReduce OpenClaw token usage and API costs through smart model routing, heartbeat optimization, budget tracking, and native 2026.2.15 features (session pruning, bootstrap size limits, cache TTL alignment). Use when token costs are high, API rate limits are being hit, or hosting multiple agents at scale. The 4 executable scripts (context_optimizer, model_router, heartbeat_optimizer, token_tracker) are local-only — no network requests, no subprocess calls, no system modifications. Reference files (PROVIDERS.md, config-patches.json) document optional multi-provider strategies that require external API keys and network access if you choose to use them. See SECURITY.md for full breakdown.
resend-cli
IncludedUse this skill when the task is specifically about operating Resend from an AI agent, terminal session, or CI job via the official resend CLI: installing/authenticating the CLI, sending/listing/updating/cancelling emails, batch sends, domains and DNS, webhooks and local listeners, inbound receiving, contacts, topics, segments, broadcasts, templates, API keys, profiles, or debugging Resend CLI/API failures. Trigger on mentions of Resend CLI, `resend`, `resend doctor`, `resend emails send`, `resend domains`, `resend webhooks listen`, `resend emails receiving`, or agent-friendly terminal automation.
alibabacloud-odps-maxframe-coding
IncludedUse this skill for MaxFrame SDK development and documentation navigation on Alibaba Cloud MaxCompute (ODPS). Helps answer MaxFrame API, concept, official example, and supported pandas API questions; create data processing programs; read/write MaxCompute tables; debug jobs (remote or local); and build custom DPE runtime images. Trigger when users mention MaxFrame, MaxCompute with MaxFrame, ODPS table processing, DPE runtime, MaxFrame docs/examples, DataFrame/Tensor operations, or GPU runtime setup. Works for both English and Chinese queries about Alibaba Cloud data processing with MaxFrame.