clawsomeflow 0.1.6

Vertical agent workflow orchestration platform on top of ClawTeam and OpenClaw

⚡ ClawsomeFlow ⚡

Make your multi-agent workflow Clawsome!

English · 简体中文

Describe your goal in natural language, and turn multi-agent collaboration into a controllable, observable, convergent engineering system.

ClawsomeFlow is a vertical-domain Agent workflow orchestration platform: define tasks as a DAG Flow, and let an async scheduler actively drive multiple Agents to collaborate in parallel, with built-in engineering guardrails such as isolation / rollback / complaint-loop / entropy management.

Full compatibility with OpenClaw, Claude Code, Codex, Cursor, Hermes and other CLI Agents.

News · Core Features · Why ClawsomeFlow · Relationship with ClawTeam · Quick Start · Roadmap

---

📰 News

• 2026-06-02: ClawsomeFlow public release 🎉

---

✨ Core Features

Building on top of ClawTeam's swarm-intelligence capabilities, ClawsomeFlow adds the two missing engineering layers — orchestration + product:

🌳 Deep OpenClaw Adaptation	🧠 AI + Precise Orchestration	🗣️ Get Everything Done in Natural Language	🔄 Complaint-Loop Mechanism
For OpenClaw's blurry session boundaries and workspace concurrency conflicts under multi-task parallelism, we apply dual session-and-directory isolation and fold recovery paths into a standard state machine.	Take control flow back from the Prompt into code: the scheduler decides dispatch, retry, timeout and convergence — behavior is controllable, with significantly fewer wasted tokens.	Flow definition, Agent creation, task orchestration, and runtime intervention can all be done in natural language via the Web UI / CLI.	A Run supports the "user complaint → reflective processing → write back experience" loop, so the system keeps self-improving.

🚀 Multi-Agent Collaboration	📊 Enterprise-Grade Observability	🔐 Isolation & Governance Together	🧩 Compatible with Existing Ecosystems
Supports OpenClaw / Claude / Codex / Cursor / Hermes collaborating in the same graph.	Every dispatch / completion / failure is recorded as a RunEvent — auditable, replayable, billable.	Three-layer isolation across team / session / worktree, avoiding cross-talk and accidental writes.	We don't reinvent the protocol layer; we reuse ClawTeam CLI + MCP and its swarm collaboration and monitoring.

🦞 The Swarm-Intelligence Foundation Inherited from ClawTeam

ClawsomeFlow stands on the shoulders of ClawTeam, faithfully inheriting its swarm-collaboration core:

• Self-organizing Agent collaboration: the Leader dispatches, Workers self-report status and results, CLI Agents are plug-and-play with no custom SDK required.
• Git Worktree workspace isolation: each Agent has an independent branch and directory, running in parallel without interference, with checkpoint / merge / cleanup support.
• Inter-Agent messaging: point-to-point inbox and broadcast, so team members share progress in real time.

On top of this, ClawsomeFlow layers in deep OpenClaw adaptation, DAG orchestration scheduling, failure convergence, human guardrails, Web productization and multi-user governance.

---

🤔 Why ClawsomeFlow?

The common pain point of multi-agent frameworks is not "insufficient model capability", but "unstable collaboration control flow": the process is written in the Prompt, the final behavior depends on the Agent's in-the-moment understanding and model quality, and the system's predictability, cost and recoverability are all too weak.

ClawsomeFlow's approach is direct: migrate coordination from natural language back into code, make concurrency isolation a default capability, and make failure handling a built-in part of the process.

🆚 Comparison with Other Agent Orchestration Platforms

Dimension	Other Multi-Agent Orchestration Platforms	✅ ClawsomeFlow
Task orchestration fit	Mostly framework-specific, bound to a single ecosystem	Task orchestration is deeply adapted to OpenClaw Agents, while also being compatible with any CLI Agent (Claude / Codex / Cursor, etc.) collaborating in the same graph
Concurrency & isolation	Easy contention in parallel, workspace conflicts, context cross-talk	Solves OpenClaw collaboration instability: workspace isolation and rollback under multi-task parallelism, and thoroughly resolves session conflicts
Control approach	Pure Prompt self-scheduling (black box) or pure code (heavy)	AI combined with precise orchestration: get everything done in natural language while the scheduler precisely controls behavior (dispatch / retry / timeout / abort)
Engineering harness	Generally missing; failures rely on Agent improvisation	Harness engineering: human checkpoints, rollbackable results, complaint-loop mechanism, periodic entropy management
Failure recovery	Relies on Agent self-healing, uncertain outcome	Clear retry / skip / abort strategies, recovery paths folded into a standard state machine
Observability	Context is mostly a black box	Full-chain RunEvent — traceable, auditable, replayable

✨ The Result?

You own the goal, and ClawsomeFlow turns multi-agent collaborative execution into a stable, controllable, convergent engineering system.

---

🧩 Relationship with ClawTeam

ClawsomeFlow is built on top of ClawTeam.

🔍 ClawTeam vs ClawsomeFlow at a Glance

Dimension	ClawTeam	ClawsomeFlow
Positioning	Swarm-intelligence protocol foundation (Agent self-organization)	Agent workflow orchestration platform
Collaboration driver	Agents self-poll and self-schedule in the Prompt	Server-side scheduler actively dispatches, deterministic execution
Task model	Kanban + dependency chain	DAG Flow compilation, Leader summarizes and converges
OpenClaw adaptation	Supported as an optional CLI Agent	Deeply adapted, resolving session and workspace concurrency conflicts
Failure & guardrails	Basic lifecycle protocol	Human checkpoints / rollback / complaint-loop / entropy management
Usage form	CLI + MCP + monitoring dashboard	Web UI + CLI, full-flow governance in natural language

---

🚀 Quick Start

Install

curl -fsSL https://clawsomeflow.com/install.sh | bash

Common Commands

# Lifecycle
csflow start
csflow stop
csflow status
csflow doctor

# Flow / Run
csflow flows list
csflow runs list
csflow runs start <flow-id> --input k=v
csflow runs abort <run-id>

# Agent governance
csflow agents list
csflow agents create "Describe the Agent you want in natural language"
csflow agents chat <agent-id> "Keep improving this Agent's capabilities"

---

🗺️ Roadmap

Phase	Content	Status
P0		🚧 In progress
P1		📅 Planned
P2		📅 Planned
P3		💡 Exploring

---

📄 License

MIT