ML PIPELINES / ZenML
ZenML MCP
Inspect ZenML projects, pipeline runs and artifacts, with optional pipeline triggers and resource mutations.
“Inspect ZenML projects, pipeline runs and artifacts, with optional pipeline triggers and resource mutations.”
01 / THE REASONING
Why this made the selection.
- Inspect ZenML projects, pipeline runs and artifacts, with optional pipeline triggers and resource mutations.
- The setup documentation makes this operating boundary explicit: The API identity governs ZenML access. Tool profile and write policy are independent; read_only removes mutation tools and pipeline triggering rather than relying on a read-only label.
02 / THE REVIEW RECORD
What we actually inspected.
Source review has boundaries.
A clear record is more useful than a “safe” badge.
Material inspected
- README.md
- LICENSE
- pyproject.toml
- server/zenml_server.py
- server/zenml_tool_catalog.py
Our findings
- The API identity governs ZenML access. Tool profile and write policy are independent; read_only removes mutation tools and pipeline triggering rather than relying on a read-only label.
- Experimental MCP Apps require Streamable HTTP. The HTTP server needs an authenticated reverse proxy for remote use: DNS-rebinding protections alone do not authenticate callers. Some log tools need deployer integrations.
Not established by this review
- No server installation, authentication, live tool execution or isolation test was performed.
- Package/container parity with the pinned source and transitive dependency security were not established.
The review applies to the material and revision named here. A newer upstream release can change its behavior.
03 / PUT IT TO WORK
Connect a server deliberately.
- Use Python 3.12 or newer and uv run /absolute/path/server/zenml_server.py from the reviewed checkout with ZENML_STORE_URL, ZENML_STORE_API_KEY, ZENML_ACTIVE_PROJECT_ID and ZENML_MCP_WRITE_POLICY=read_only for inspection.
- Verify the connected account, available tools and target resource before authorizing mutations.
Before you start
- Pinned package Python requirement: >=3.12.
- A compatible local process MCP client.
- The API identity governs ZenML access. Tool profile and write policy are independent; read_only removes mutation tools and pipeline triggering rather than relying on a read-only label.
04 / FOLLOW THE EVIDENCE
The source trail.
Our notes are separate from the original resource.
Check upstream before adopting a new version.
https://github.com/zenml-io/mcp-zenml/blob/693ac197b4013a2b70971464f78bdf3211ef22c0/README.md
Supports: description, setup, capabilities, review, limitations
https://github.com/zenml-io/mcp-zenml/blob/693ac197b4013a2b70971464f78bdf3211ef22c0/LICENSE
Supports: license
- Pinned package metadata ↗Checked
https://github.com/zenml-io/mcp-zenml/blob/693ac197b4013a2b70971464f78bdf3211ef22c0/pyproject.toml
Supports: install, language, review
https://github.com/zenml-io/mcp-zenml/blob/693ac197b4013a2b70971464f78bdf3211ef22c0/server/zenml_server.py
Supports: review, install, access, limitations
https://github.com/zenml-io/mcp-zenml/blob/693ac197b4013a2b70971464f78bdf3211ef22c0/server/zenml_tool_catalog.py
Supports: review, install, access, limitations