---
title: "Regularized Recursive Self-Improvement of Agent Harnesses (RRSI) by Google Research | AIDeveloper44"
url: https://aideveloper44.com/product/regularized-recursive-self-improvement-of-agent-harnesses-rrsi-6ab9eaa741e249e356902c71
description: "RRSI is a Google Research framework for regularized recursive self-improvement of LLM agent harnesses, enhancing performance across coding and engineering tasks."
lang: en
---

← all releases (https://aideveloper44.com/releases)

Google Research · Sep 27, 2026 · rolling · open source

# Regularized Recursive Self-Improvement of Agent Harnesses (RRSI)

RRSI is a framework for the recursive self-improvement of LLM agent harnesses, designed to optimize prompts, control flow, tools, and memory. It addresses overfitting in agent evolution by regularizing both the proposal and selection processes, ensuring robust performance across diverse task domains.

AI Agents Recursive Self-Improvement LLM Optimization Automated Engineering

visit website: https://regularized-rsi.com/

github ↗: https://github.com/google-research/rrsi

overview

key features

- 01 Regularized search trajectory to prevent overfitting during agent evolution
- 02 Open edit space allowing modifications to prompts, tools, memory, and control flow
- 03 Git-based candidate management for auditability and version control
- 04 Multi-domain support including coding, workspace, and engineering design
- 05 Automated critic and analyst components for screening and evaluating harness edits

related products browse all → (https://aideveloper44.com/releases)

MiroThinker open source MiroThinker is an open-source search agent model developed by MiroMindAI, designed for tool-augmented reasoning and real-world information seeking. It aims to match the deep research capabilities of leading AI models like OpenAI's Deep Research and Google's Gemini Deep Research.: https://aideveloper44.com/product/mirothinker-69b8df317e0f5aa1e309bd18
Letta Code SDK open source The Letta Code SDK is a software development kit that enables developers to build deeply personalized agents with persistent memory that learn over time. It serves as the interface to Letta Code, facilitating the creation of stateful agents capable of continuous learning and improvement.: https://aideveloper44.com/product/letta-code-sdk-697a7e96c0c68ce2ebf09661
OB-1 open source OB-1 is a self-improving coding agent developed by OpenBlock Labs, designed to autonomously handle the full development lifecycle, from project management to pull requests. It integrates seamlessly into existing workflows, enhancing productivity and code quality.: https://aideveloper44.com/product/ob-1-69b316a10797d7961e69d575

## Structured data

```json
[
  {
    "@context": "https://schema.org",
    "@type": [
      "WebSite",
      "NewsArticle"
    ],
    "name": "AIDeveloper44",
    "url": "https://aideveloper44.com",
    "isAccessibleForFree": true,
    "isPartOf": {
      "productID": "CAowsfXgCw:openaccess",
      "@type": [
        "Product"
      ]
    }
  },
  {
    "@context": "https://schema.org",
    "@type": "Organization",
    "name": "AIDeveloper44",
    "url": "https://aideveloper44.com",
    "logo": "https://qtrypzzcjebvfcihiynt.supabase.co/storage/v1/object/public/base44-prod/public/696b2bd107af20deac4cc200/9f77c0b91_a-digital-illustration-of-a-white-llama-_j4dyH5xoRnKf_3rtw-mvEQ_OdqL2BrSTtKHjJAuhTp_zg.png"
  },
  {
    "@context": "https://schema.org",
    "@type": "BreadcrumbList",
    "itemListElement": [
      {
        "@type": "ListItem",
        "position": 1,
        "name": "Home",
        "item": "https://aideveloper44.com/"
      },
      {
        "@type": "ListItem",
        "position": 2,
        "name": "product",
        "item": "https://aideveloper44.com/product"
      },
      {
        "@type": "ListItem",
        "position": 3,
        "name": "Product Detail",
        "item": "https://aideveloper44.com/product/regularized-recursive-self-improvement-of-agent-harnesses-rrsi-6ab9eaa741e249e356902c71"
      }
    ]
  },
  {
    "@context": "https://schema.org",
    "@type": "SoftwareApplication",
    "name": "Regularized Recursive Self-Improvement of Agent Harnesses (RRSI)",
    "description": "RRSI is a framework for the recursive self-improvement of LLM agent harnesses, designed to optimize prompts, control flow, tools, and memory. It addresses overfitting in agent evolution by regularizing both the proposal and selection processes, ensuring robust performance across diverse task domains.",
    "provider": {
      "@type": "Organization",
      "name": "Google Research"
    },
    "offers": {
      "@type": "Offer",
      "price": "Free & Open source"
    }
  },
  {
    "@context": "https://schema.org",
    "@type": "Organization",
    "name": "AINews.sh",
    "alternateName": "AI News",
    "url": "https://ai-dev-pulse-ac4cc200.base44.app",
    "logo": "https://ai-dev-pulse-ac4cc200.base44.app/logo.png",
    "description": "Developer-focused AI product releases, reviews, and news. Your source for the latest AI tools, open-source models, and platform updates.",
    "foundingDate": "2026",
    "slogan": "Stay ahead with AI product intelligence",
    "sameAs": [],
    "knowsAbout": [
      "Artificial Intelligence",
      "Machine Learning",
      "Large Language Models",
      "AI Tools",
      "Open Source AI",
      "AI Frameworks",
      "Developer Tools"
    ]
  }
]
```