🧠 Make your agents learn from experience. Now available as a hosted solution at kayba.ai
The living ecosystem where AI agents complete tasks through workflow loops, improve through iterative execution, are evaluated by mentor agents or humans in the loop, and turn completed work into reusable work experience and data to improve future agents.
Search, understand, reproduce, and improve an idea with ease
🪞 Make your agents recursively self-improve
This repository hosts a customized PPO based agent for Carla. The goal of this project is to make it easier to interact with and experiment in Carla with reinforcement learning based agents -- this, by wrapping Carla in a gym like environment that can handle custom reward functions, custom debug output, etc.
[NAACL 2025] KnowAgent: Knowledge-Augmented Planning for LLM-Based Agents
[ACL 2024] AutoAct: Automatic Agent Learning from Scratch for QA via Self-Planning
Self improving agents through iterations
[ACL 2024] Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View
Reflexion episodic failure memory buffer accumulating self-critiques and trial corrective plans
Reflexion episodic failure memory buffer accumulating self-critiques and trial corrective plans
The lightest CLI agent for Python. No heavy deps, no complex chains. pip install, set your API key, and run AI tasks anywhere.
Codex plugin that exports local Codex learning signals into Hermes Agent
Local-first trajectory learning for coding agents. Every run makes the next one better.
Agent Learning Protocol — the fourth pillar of agent-native content. An open standard for structured learning pathways that AI agents can follow. Like MCP for learning.
A PyTorch re-implementation of World Models (Ha & Schmidhuber, 2018) for CarRacing-v3. The agent solves the track by "dreaming"—using a VAE for perception, an MDN-RNN for memory, and CMA-ES for controller evolution.
A 30-day AI Agent learning plan for engineers — from agent loops & context engineering to memory/RAG, tools/MCP, evals and multi-agent collaboration. ~60 min/day, ending with a runnable, evaluable mini agent.
Local Codex skills and automations for reusable agent learnings
To associate your repository with the agent-learning topic, visit your repo's landing page and select "manage topics."