UMLCAUTION
UML-MCP Server is a UML diagram generation tool based on MCP (Model Context Protocol), which can help users generate various types of UML diagrams through natural language description or directly writing PlantUML and Mermaid and Kroki https://uml-mcp.vercel.app/mcp
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://github.com/antoinebou12/uml-mcp/actions/workflows/test.yml) [](https://github.com/antoinebou12/uml-mcp/actions/workflows/build.yml) [](https://github.com/antoinebou12/uml-mcp/actions/workflows/docs.yml) [](https://github.com/antoinebou12/uml-mcp/actions/workflows/deploy.yml) [](https://github.com/antoinebou12/uml-mcp/releases) [](https://github.com/antoinebou12/uml-mcp/stargazers) [](https://github.com/antoinebou12/uml-mcp/network/members) [](https://github.com/antoinebou12/uml-mcp/issues) [](https://opensource.org/licenses/MIT) [](https://www.python.org/downloads/) [](https://github.com/astral-sh/ruff) [](https://github.com/astral-sh/uv) [](https://modelcontextprotocol.io/) [](https://uml-mcp.vercel.app/mcp) [
d0649ecc146eOBSERVED · 2026-10-07Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add uml-mcp-console --env MCP_AUTH_RESOURCE_URL=${MCP_AUTH_RESOURCE_URL} --env MCP_LINT_TOKEN=${MCP_LINT_TOKEN} --env MCP_SMOKE_TOKEN=${MCP_SMOKE_TOKEN} -- npx -y [email protected]{
"mcpServers": {
"uml-mcp-console": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"MCP_AUTH_RESOURCE_URL": "${MCP_AUTH_RESOURCE_URL}",
"MCP_LINT_TOKEN": "${MCP_LINT_TOKEN}",
"MCP_SMOKE_TOKEN": "${MCP_SMOKE_TOKEN}"
}
}
}
}Exposed tools (2)
1 read · 1 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
generate_uml_url | write | Return a Kroki render URL for a diagram (POST /kroki_encode). |
list_diagram_types | read | List supported diagram types and output formats (GET /supported_formats). |
Trust audit
CAUTIONgrade C · trust 78/100 Install with care. The audit found things worth knowing before you trust its output.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | WARN |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (4 observation(s))
- Network
- declared (7 observation(s))
- Shell
- declared (1 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- found
Findings (25)
return __import__(module_name)
CMD curl -fsS http://127.0.0.1:8000/health || exit 1
TOKEN = "kroki-setup-token-0123456789"
TOKEN = "test-setup-token-0123456789"
.pre-commit-config.yaml
.vercelignore
.helmignore
UML_MCP_TEST_KROKI_URL: http://127.0.0.1:8001
if curl --fail --silent http://127.0.0.1:8766/health | jq -e '.status == "healthy"' >/dev/null; then
--url http://127.0.0.1:8766/mcp \
Docs locally: `uv run mkdocs serve` → http://127.0.0.1:8000
const bytes = Uint8Array.from(atob(result.content_base64), (c) => c.charCodeAt(0));
check_svg(base64.b64decode(encoded), labels)
png = base64.b64decode(images[0].data) if images else b""
png = base64.b64decode(out.get("content_base64") or "")check_svg(base64.b64decode(encoded))
@radix-ui/react-dialog, @radix-ui/react-dropdown-menu, @radix-ui/react-label, @radix-ui/react-select, @radix-ui/react-separator, @radix-ui/react-slot, @radix-ui/react-switch, @radix-ui/react-tabs
2. **Never forward a client's token.** The MCP server is the audience of the
curl -s -X POST https://mcp.contoso.com/mcp -H "Authorization: Bearer $TOKEN" \
After creating the server, open it on Smithery → **Settings** → **Session configuration** (or equivalent) and paste or upload the contents of `smithery-config-schema.json`. See [Smithery Session Confi
public/ressources/formal-13-12-09.pdf
public/ressources/kroki_cheatsheet_20210515_v1.1_EN.pdf
| `MCP.Read.All` | Applications | read (client credentials) |
| `MCP.Write.All` | Applications | read + write (client credentials) |
1. **Use Node 20+** when running the CLI (e.g. install Node 20 or use [nvm](https://github.com/nvm-sh/nvm): `nvm use 20` then run the `deploy` command again).
Gates applied: no_behavioural_pass.
d0649ecc146efull audit observations/trust-audit/mcp-server/antoinebou12__uml-2.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | d0649ecc146e | CAUTION | C | 78 | first audit |
Questions
What is the UML MCP server?
UML-MCP Server is a UML diagram generation tool based on MCP (Model Context Protocol), which can help users generate various types of UML diagrams through natural language description or directly writing PlantUML and Mermaid and Kroki https://uml-mcp.vercel.app/mcp
What tools does UML expose?
2 in total: 1 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is UML safe to connect to an agent?
With care. The audit graded it C (78/100) and found 25 things worth knowing before you trust this server, listed below with the exact line each was found on.
What credentials does UML need?
It reads MCP_AUTH_RESOURCE_URL, MCP_LINT_TOKEN and MCP_SMOKE_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does UML run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as uml-mcp-console at 1.4.0.
How current is this page?
The grade is for one exact copy of the source (d0649ecc146e), read on 2026-10-07. The repository is watched and re-audited when it changes.