MCP (Model Context Protocol) servers that expose Systemonomic's Work Domain Analysis, ATSS assessment, and organizational design capabilities to AI agents (Claude Desktop, Cursor, Claude Code, etc.).
pip install systemonomic-mcpLog in to Systemonomic
Go to Profile → API Keys
Click Generate API Key
Copy the key (starts with sk_sys_) — it's shown only once
Set the environment variable:
export SYSTEMONOMIC_API_KEY="sk_sys_your_key_here"Optionally, point to a different API endpoint (defaults to production):
export SYSTEMONOMIC_API_URL="https://your-dev-backend.up.railway.app"Add to your claude_desktop_config.json:
{
"mcpServers": {
"systemonomic-wda": {
"command": "python",
"args": ["-m", "systemonomic_mcp.wda_server"],
"env": {
"SYSTEMONOMIC_API_KEY": "sk_sys_your_key_here"
}
},
"systemonomic-atss": {
"command": "python",
"args": ["-m", "systemonomic_mcp.atss_server"],
"env": {
"SYSTEMONOMIC_API_KEY": "sk_sys_your_key_here"
}
},
"systemonomic-org": {
"command": "python",
"args": ["-m", "systemonomic_mcp.org_server"],
"env": {
"SYSTEMONOMIC_API_KEY": "sk_sys_your_key_here"
}
}
}
}In Cursor Settings → MCP Servers, add each server:
Name: systemonomic-wda
Command: python -m systemonomic_mcp.wda_server
Environment: SYSTEMONOMIC_API_KEY=sk_sys_...
Repeat for atss_server and org_server.
Related MCP server: systems-thinking-mcp
systemonomic-wda — Work Domain AnalysisTool | Description |
| List all projects |
| Get complete project state |
| Create a new project |
| Get project statistics |
| List WDA nodes |
| Create a node at a WDA level |
| Update a node's label/level/description |
| Delete a node |
| List means-ends links |
| Link two nodes |
| Remove a link |
| AI-generate a full WDA from a text description |
| Export project as JSON |
| Import nodes and links |
systemonomic-atss — Assessment & TasksTool | Description |
| List project tasks |
| Create a task |
| Auto-generate tasks from WDA Objects |
| AI-derived task suggestions |
| List pending suggestions |
| Promote suggestions to tasks |
| Run ATSS assessment on all tasks |
| Get stored assessment results |
| Save assessment results |
| List assessment run history |
systemonomic-org — Organizational DesignTool | Description |
| Get current roles, org units, allocations |
| Save org design |
| AI-generated restructure proposal |
| Apply a restructure proposal |
| Validate RACI matrix |
| Create version snapshot |
| List snapshots |
| Export as JSON |
| Generate comprehensive PDF report |
| Check report readiness |
You: Generate a WDA for our university procurement department. They handle purchase requests, vendor management, contract negotiation, and compliance with government regulations.
Claude: Uses
create_project→generate_wda→ returns the generated hierarchy
You: For the procurement project, derive tasks from the WDA and run an automation assessment.
Claude: Uses
generate_tasks_from_wda→run_atss_batch→ summarizes automation candidates
You: Create a PDF report for the procurement project.
Claude: Uses
generate_pdf_report→ saves the PDF
# Run a server locally for testing
cd mcp
pip install -e .
SYSTEMONOMIC_API_KEY=sk_sys_... python -m systemonomic_mcp.wda_server
# Use the MCP inspector
SYSTEMONOMIC_API_KEY=sk_sys_... mcp dev src/systemonomic_mcp/wda_server.pyAccept task suggestions, promoting them to actual project tasks.
Args: project_id: The project containing the suggestions suggestion_ids: List of suggestion IDs to accept
Inputs: project_id (required), suggestion_ids (required). Returns: result
Create a new task in a project.
Args: project_id: The project to add the task to name: Task name description: Optional task description mode: One of: manual, semi-auto, auto (default: manual)
Inputs: project_id (required), name (required), description, mode. Returns: result
Use AI to derive detailed task suggestions from WDA objects.
More sophisticated than generate_tasks_from_wda — uses an LLM to analyze each WDA object and suggest tasks with descriptions.
Args: project_id: The project to analyze provider: LLM provider — gemini, claude, or openai (default: gemini)
Inputs: project_id (required), provider. Returns: result
Auto-generate tasks from the WDA Objects level.
Analyzes the Objects (lowest level) of the WDA and creates corresponding control tasks. This is the standard first step before running ATSS.
Inputs: project_id (required). Returns: result
Get stored ATSS results for a project.
Returns previously persisted assessment results, including scores, classifications, and reasoning for each task.
Inputs: project_id (required). Returns: result
List all ATSS assessment runs for a project, with timestamps and summaries.
Inputs: project_id (required). Returns: result
List all pending task suggestions for a project.
Suggestions are AI-generated task proposals that haven't been accepted yet.
Inputs: project_id (required). Returns: result
List all tasks in a project.
Each task has an id, name, description, mode (manual/semi-auto/auto), and links to WDA nodes.
Inputs: project_id (required). Returns: result
Persist ATSS assessment results to the project.
Args: project_id: The project to save results to rows: List of ATSS result objects (from run_atss_batch output)
Inputs: project_id (required), rows (required). Returns: result
Run ATSS (Automated Task Suitability Scoring) on all tasks in a project.
Each task is assessed across multiple gates (data availability, rule-base, exception handling, etc.) and scored 0-100 for automation suitability.
Args: project_id: The project whose tasks to assess provider: LLM provider — gemini, claude, or openai (default: gemini) model: Specific model name (optional, uses provider default)
Returns scored results for each task with classification (Automate / Augment / Manual) and reasoning.
Inputs: project_id (required), provider, model. Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 10 tools
Most tools have distinct purposes, but there is some potential confusion between 'derive_task_suggestions' and 'generate_tasks_from_wda' as both generate tasks from WDA objects, with the former being described as 'more sophisticated.' The other tools are clearly differentiated, covering suggestion management, task creation, ATSS assessment, and listing operations.
All tool names follow a consistent snake_case pattern with clear verb_noun structures (e.g., 'accept_suggestions', 'create_task', 'list_tasks'). This uniformity makes the tool set predictable and easy to navigate, with no deviations in naming conventions.
With 10 tools, the server is well-scoped for managing tasks, suggestions, and ATSS assessments in a project automation system. Each tool serves a specific role in the workflow, from creation and listing to analysis and persistence, without feeling bloated or sparse.
The tool set covers core workflows for task and suggestion management, including creation, listing, and assessment, but lacks update and delete operations for tasks or suggestions. This minor gap might require workarounds, but the overall surface supports the domain of project automation and AI-driven task analysis effectively.
Task-by-task AI exposure assessment of a job (Spain): 0-100 index, task map, weekly plan.
Task and planning workspace for humans collaborating with AI agents
Accountable AI workforce workflows with jobs, skills, rules, evidence, and mentoring.
Provides UX capabilities to enhance the design output and understanding of AI systems.
