A Model Context Protocol (MCP) server for AI agent workflow orchestration. Create structured, fidelity-enforced workflows that agents discover, navigate, and execute to fulfill user goals.
Docs Site • Architecture • API • Workflow Fidelity • Development • Workflows
Workflow Server guides AI agents through structured, multi-step workflows. A single always-applied IDE rule bootstraps the agent — from there, the server handles workflow discovery, session management, and step-by-step navigation.
- Discover — The agent learns which workflows exist and how to begin
- Start — A session is started for the matched workflow
- Navigate — The agent moves through activities in order, loading each phase’s steps and guidance as needed
- Execute — Work proceeds activity by activity, pausing at checkpoints for user decisions and following transitions between phases
User Goal → Workflow → Activities → Techniques → Tools
- Workflows — define the overall process (e.g., implement a feature from issue to merged PR)
- Activities — are phases within a workflow (e.g., plan, implement, review, validate)
- Techniques — are markdown definitions of a capability, with optional rules
- Tools — are the operations the agent invokes
This project is tailored towards the use of the Cursor IDE. To use with Claude Code it is recommended to setup the Claude Code plugin for Cursor.
See: setup.md for detailed setup instructions (including Cursor workspace at §3).
With the server connected and target project initialised, tell the agent in your chat session what you want to do, for example:
Start a new work-package workflow for Issue #1000
Resume the work-package workflow for PR #1000
The agent matches the request to the appropriate activity and guides you through the structured phases.
The server registers 17 MCP tools across five concerns. See docs/api-reference.md for full signatures.
MIT License - see LICENSE for details.