Breaking

Tom Dörr releases The-Claude-Protocol to orchestrate multiple Claude Code agents via Beads CLI

July 2, 2026 at 15:57 EDT

AI Agents
Software Dev & Coding
Open Source

Developer Tom Dörr has released The-Claude-Protocol, a multi-agent orchestration framework for Anthropic's CLI-based AI coding agent Claude Code. Built around the Beads CLI, it automates task management to run long, complex coding work in parallel.

v0.6.0 · Maestro · MIT License

An Orchestration Layer That Conducts Claude Code

Maestro coordinates specialized agents, modular Skills, event-driven Hooks and persistent memory to automate AI-driven development on Anthropic's Claude Code CLI — architecture before implementation, tests before code.

211

GitHub stars

12+

Modular Skills

80%

Target test coverage

Releases · 121 commits

How a task flows through Maestro

COMMAND

/maestro <task>

You describe the goal

→

ORCHESTRATE

grandmaster

Decomposes & delegates

→

AGENTS + SKILLS

TDD · design · debug

Tests first, then build

→

AUTONOMOUS QA

Ralph Wiggum

Verify on evidence

Long-Term Memory writes stack, decisions, errors & history to brain.jsonl — context persists across sessions.

Ralph Wiggum · four pillars of autonomous QA

Completion is judged on evidence, not assumption.

Proactive Gate

Blocks work until intent is clear

Reflection Loop

Self-reviews each iteration

Verification Matrix

Targets 80% test coverage

Circuit Breaker

Detects and halts stalls

What supporters cite

Runs multiple sessions in parallel
Prevents context loss between runs
Strict control layers cut debugging rework

Open questions

Single-provider dependence risks policy shifts
Node.js + Claude Code CLI adds setup overhead
Recent release — real-world track record still limited

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.