What is the difference between Working and Episodic memory?

Working memory caches immediate codebase facts (like helper methods) for all agents to share, while Episodic memory stores long-term learnings from past tasks to improve future performance on similar problems.

Can I use these patterns for TDD?

Yes, the skill includes specialized LOOP and CONDITIONAL patterns designed specifically for iterative Test-Driven Development cycles, running tests and fixing code until they pass.

What is the ReAcTree framework?

ReAcTree is a hierarchical LLM agent framework that uses tree structures and control flow nodes to solve complex, long-horizon tasks more effectively than traditional single-agent models.

How does this skill improve development speed?

It uses Parallel execution nodes to process independent tasks simultaneously—such as building models, services, and tests at once—reducing total workflow time by up to 50%.

ReAcTree Agent Coordination Patterns

Name: ReAcTree Agent Coordination Patterns
Author: Kaakati

byKaakati

•

Productivity & Workflow

Orchestrates complex development tasks using hierarchical agent trees, sophisticated control flow logic, and dual memory systems.

Based on the 2024 ReAcTree research paper, this skill enables Claude to move beyond simple sequential reasoning by implementing hierarchical agent structures. It utilizes specialized control flow nodes—including Parallel, Fallback, LOOP, and CONDITIONAL—to decompose long-horizon Rails development tasks into manageable, high-success workflows. By integrating Working Memory for shared codebase facts and Episodic Memory for learning from past successes, it significantly increases task completion rates and reduces redundant analysis operations.

Key Features

01Research-proven 97% improvement in success rates over monolithic ReAct approaches

02Parallel execution support to accelerate development workflows by 30-50%

03Hierarchical task decomposition using behavior tree-inspired control flow nodes

04Iterative LOOP and CONDITIONAL branching for automated TDD and error recovery

052 GitHub stars

06Dual memory systems (Working & Episodic) for caching facts and learning from past episodes

Use Cases

01Creating resilient deployment and integration scripts with automated fallback and retry logic

02Automating TDD cycles with iterative red-green-refactor loops and automated verification

03Accelerating complex feature implementation by running independent service and component development in parallel

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kaakati/rails-enterprise-dev reactree-patterns

For use in Claude.ai and ChatGPT

Download Skill