Skip to main content
Glama

OSOP MCP Server

License Python OSOP Compatible

MCP (Model Context Protocol) server that exposes OSOP workflow operations as tools for AI agents. Any MCP-compatible client (Claude Desktop, Claude Code, OpenClaw, Cursor, etc.) can validate, run, render, test, optimize, and assess security risks of OSOP workflows.

Website: osop.ai | Editor: osop-editor.vercel.app

Tools

Tool

Description

osop.validate

Validate an .osop.yaml file against the OSOP schema. Returns errors and warnings.

osop.risk_assess

Analyze workflow for security risks — permission gaps, missing approval gates, destructive commands, cost exposure. Returns risk score (0-100) and actionable findings.

osop.run

Execute CLI/API nodes in a workflow with timeout support. Supports dry-run mode.

osop.render

Render a workflow as Mermaid or ASCII diagram.

osop.test

Run test cases defined in the workflow and report pass/fail results.

osop.optimize

Analyze a workflow for redundancies, parallelization opportunities, and bottlenecks.

osop.report

Generate HTML or text reports from workflow + optional execution log.

osop.convert

Unified format conversion — import/export between OSOP and 6 external formats.

osop.import

Convert external formats (CrewAI, n8n, GitHub Actions, Airflow, Argo, LangGraph) into OSOP.

osop.export

Convert an OSOP workflow to CrewAI, n8n, or Argo format.

osop.diff

Structural comparison of two OSOP workflows — added/removed/changed nodes and edges.

osop.notion

Convert OSOP workflow to Notion database-ready structures (schema + page payloads).

Installation

pip install osop-mcp

Usage

Standalone

python -m osop_mcp

The server listens on stdio by default (MCP standard transport).

Docker

docker build -t osop-mcp .
docker run -i osop-mcp

Claude Desktop / Claude Code

Add to your claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%/Claude/):

{
  "mcpServers": {
    "osop": {
      "command": "python",
      "args": ["-m", "osop_mcp"],
      "env": {}
    }
  }
}

For Claude Code, add to your project's .mcp.json:

{
  "mcpServers": {
    "osop": {
      "command": "python",
      "args": ["-m", "osop_mcp"]
    }
  }
}

Docker-based Configuration

{
  "mcpServers": {
    "osop": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "osop-mcp"]
    }
  }
}

Quick Start: Risk Assessment

Once installed, ask your AI agent:

"Analyze the security risks of my deployment workflow"

The agent will use osop.risk_assess to:

  1. Walk the workflow DAG and identify high-risk nodes

  2. Check for missing approval gates before destructive operations

  3. Flag overly broad permissions (write:*, admin:*)

  4. Detect destructive CLI commands (rm -rf, kubectl delete, terraform destroy)

  5. Calculate cost exposure from agent nodes

  6. Return a risk score (0-100) with verdict: safe / caution / warning / danger

Example output:

{
  "overall_score": 72,
  "verdict": "danger",
  "total_findings": 5,
  "findings": [
    {
      "severity": "critical",
      "title": "CRITICAL risk node without approval gate",
      "node_id": "deploy-prod",
      "suggestion": "Add approval_gate with required: true before this node."
    }
  ]
}

Environment Variables

Variable

Description

Default

OSOP_MCP_TRANSPORT

Transport protocol (stdio or sse)

stdio

OSOP_MCP_PORT

Port for SSE transport

8080

OSOP_LOG_LEVEL

Logging level

INFO

Development

git clone https://github.com/Archie0125/osop-mcp.git
cd osop-mcp
pip install -e ".[dev]"
pytest

What is OSOP?

OSOP (Open Standard Operating Protocol) is the OpenAPI of workflows. It standardizes how workflows, SOPs, and automation pipelines are defined, validated, and executed — across AI agents, CI/CD tools, and enterprise processes.

License

Apache License 2.0 — see LICENSE for details.

-
security - not tested
F
license - not found
-
quality - not tested

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Archie0125/osop-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server