ProcessDesignAgents — DEPRECATED

⚠️ This repository is archived. The active project has moved to the process-orchestrator-plugin Agent Plugin.

What Changed

The multi-agent process design workflow has been repackaged as a portable Agent Plugins v1.0.0 plugin. Instead of a Python/LangGraph codebase, the pipeline is now defined as 13 SKILL.md files (orchestrator + 12 sub-agents) with a shared state schema, PES MCP integration, and an OpenClaw extension namespace.

Old Architecture (this repo)

• Python LangGraph state machine
• LangChain prompt templates
• CLI via python -m cli.main
• Tied to Python runtime

New Architecture (process-orchestrator-plugin)

• Portable Agent Plugins format (client-agnostic)
• 13 SKILL.md files with YAML frontmatter
• Pipeline defined in ai.openclaw/config/graph.json
• PES MCP server for calculations (Streamable HTTP + stdio)
• Works in any conformant client (OpenClaw, Cursor, VS Code, etc.)

Migration

If you were using this repo, the new plugin is at:

https://github.com/may3rd/process-orchestrator-plugin

The 12-step pipeline, state schema, and agent responsibilities are preserved. The new format is simpler, more portable, and requires no Python runtime to define the workflow.

License

MIT License (see LICENSE for details).