Framework for evaluating and improving agents

Multi-Agent Self-Supervision

Official Repo for "HarnessTTT: Learn to Optimize Harness at Test-Time for Scientific Discovery"

DiscoveryRSI: Recursively Learning to Discover Valuable Hypotheses

SciLaws-Bench: Can LLMs Discover Scientific Laws in Real and Parallel Worlds?

A curated list for Self-Improvement in Foundation Model Based Agentic Systems.

SHELDON is a reasoning harness that searches observations and hypotheses as two separate spaces, focusing on finding unknown observations to generate new hypotheses.

Evolving what multimodal agents see and how they read it. Framework, ten reusable task skills, and paper resources.

SmolWAM: a 0.88B latent world-action model for robot manipulation on edge devices. NeurIPS 2026 Workshop on Robotics World Modeling.

WikiSkill (arXiv:2608.27454) for Hermes Agent — self-evolving agent skills via a persistent knowledge wiki. Faithful Algorithm 1 implementation with real agent runs, isolated skill gating, and a do…

EvoSkill — An open-source framework that automatically discovers and synthesizes reusable agent skills from failed trajectories to improve coding agent performance.

SpecEvo: Speculative Evolution with Large Language Models for Cost-Efficient Scientific Discovery

A curated list of academic papers and resources on Physical AI — focusing on Vision-Language-Action (VLA) models, world models, embodied ai, and robotic foundation models.

A comprehensive list of papers about dual-system VLA models, including papers, codes, and related websites.

Autonomous research system for measurable, computer-executable research.

LLM-ACES: Closed-Loop Discovery of Dynamical Systems with LLM-Guided Adaptive Search

[AAAI 2026 Oral] MOTIF: Multi-strategy Optimization via Turn-based Interactive Framework

A low-cost and high-performance evolutionary engine for code and prompt optimization

SpecEvo: Speculative Evolution with Large Language Models for Cost-Efficient Scientific Discovery

AI-Driven Scientific, Algorithmic, and Systems Discovery

CORAL is a robust, lightweight infrastructure for multi-agent autonomous self-evolution, built for autoresearch.

Unlimited FREE AI coding. Connect Claude Code, Codex, Cursor, Cline, Copilot, Antigravity to FREE Claude/GPT/Gemini via 40+ providers. Auto-fallback, RTK -40% tokens, never hit limits.

"OpenSpace: The Skill Management Layer for AI Agents" -- https://open-space.cloud/

Open-source (Apache-2.0) orchestrator for Claude Code, Codex & 30+ coding agents. Plan, run and supervise agent teams from task to merged PR. Desktop (macOS, Windows, Linux) + iPhone & Android apps.