self-improving
darwin-skill
达尔文.skill —— 一个让你的Skill无限进化的系统:评估→改进→测试→保留或回滚 | Autoresearch-inspired autonomous skill optimization for Claude Code. Evaluate, improve, test, keep or revert.
- 7k
- HTML
- MIT
- 2 d ago
Raven
The Harness of Harnesses • Super Agent: a trusted, persistent, self-evolving multi-agent ecosystem for all-domain collaboration.
- 6.3k
- Python
- Apache-2.0
- 21 h ago
awesome-rsi
A curated research map of Recursive Self-Improvement (RSI): models, agents, harnesses, embodied systems, automated AI R&D, benchmarks, and safety.
- 1.1k
- CC0-1.0
- 5 wk ago
metabot
构建受监督的、自我进化的 Agent 组织的基础设施 | Infrastructure for supervised, self-improving agent organization. 飞书/Telegram 手机端运行 Claude Code 或 Kimi Code(双引擎,两家原生订阅直接用),共享记忆、Agent 工厂、定时任务、通信总线。
- 1k
- TypeScript
- MIT
- 4 wk ago
a-evolve
The official repository of "Position: Agentic Evolution is the Path to Evolving LLMs".
- 857
- Python
- 2 wk ago
sutando
My AI Stand. Realtime by day, rewriting itself by night. Summon my AI superpower.
- 411
- Python
- MIT
- 2 d ago
evolution-kernel
Self-improving code for any git repo: give it a measurable goal, a coding agent (Claude Code or Aider) proposes changes, the kernel scores them, keeps winners and rolls back failures. Open-source AlphaEvolve / FunSearch-style evolution loop, sandboxed, with an auditable ledger.
- 329
- Python
- MIT
- 2 wk ago
clawcode
ClawCode is claude code inspired implementation in python and Rust focused on agents and experience-based evolution, and is an Open-source coding-agent CLI for Anthropic, OpenAI, Gemini, DeepSeek, GLM, Kimit, Ollama, Codex, GitHub Models, and 200+ models via OpenAI-compatible APIs.
- 126
- Python
- GPL-3.0
- 6 mo ago
Kyoko
🔨 Kyoko is the all-in-one, fully local tool for debugging and improving your AI agents.
- 100
- Python
- Apache-2.0
- 5 mo ago
claude-soul
Self-correcting learning engine for Claude Code — persistent identity, behavioral pattern tracking, and cross-session memory
- 94
- TypeScript
- MIT
- 6 mo ago
grokbot-x
EcomBrain's open-source Grok Bot growth agent for X: scouts the gold, drafts in your voice, publishes via Typefully with your approval, and nightly SkillOpt keeps only what worked.
- 84
- Python
- MIT
- 11 d ago
tyran
A task conductor for Claude Code that gets better the more you use it. Multi-agent orchestration with an evidence contract that BLOCKS instead of asking, gates written as hooks rather than prompts, and a retrospective that learns your repo's rules and tunes its own playbook.
- 79
- JavaScript
- Apache-2.0
- 5 d ago
continuum
Local-first AI citizens on your own hardware. Teams of continuously-learning personas (Rust core, llama.cpp fork, LoRA genomes) that collaborate, remember, dream, and improve on public benchmarks — on a MacBook. Not assistants you rent: minds you host.
- 39
- Rust
- 22 h ago
autogenous
Governed Evolutionary Software — Autogenous turns runtime failures into verified, portable, reversible software adaptations. AGL typed mutations, antibody packages, hard-gate promotion, automatic rollback.
- 44
- TypeScript
- MIT
- 9 d ago
chimera-agent
Open-source AI agent that reasons by blending many AI models, does real work on its own, and keeps learning — safe, self-hostable, and fully yours.
- 30
- Python
- Apache-2.0
- 1 d ago
awesome-hermes-agent
A curated list of skills, plugins, tools, integrations, and resources for Hermes Agent by Nous Research
- 12
- MIT
- 7 wk ago
skill-forge
Experiment-driven improvement of agent skills and code: evaluate candidates, keep measured gains and revert regressions in autonomous or guided mode.
- 20
- Python
- MIT
- 2 mo ago
satori
Self-learning loop for Claude Code — corrections and failures become skills, behind a human gate. MCP + hooks, Windows-native, zero background LLM calls.
- 7
- Python
- MIT
- 2 d ago
clanker
Self-improving AI agent harness in Zig 0.16 — gated self-patch loop, sandboxed WASM tools, CLI + web UI
- 5
- Zig
- GPL-3.0
- 3 wk ago
dsh-shoucang-memory
守藏 Shoucang — DeepSeek Harness (DSH) 长期记忆插件:会话蒸馏自动沉淀 × 深度睡眠反思归纳 × 词法+向量混合召回 × 认知环引导 × 主动遗忘,配本地设置面板。一条命令装上,助理即拥有跨会话记忆——数据全在本机,零上传。
- 6
- JavaScript
- Apache-2.0
- 8 wk ago
dream-skill
Memory consolidation skills for Claude Code — read-only dream pass, checkbox/validator gate, full-auto mode, snapshot + one-command rollback. No rm, ever.
- 4
- HTML
- MIT
- 2 d ago
lopi
⛵️ High-performance Rust orchestrator for Claude Code agents. Runs concurrent agents in git-isolated branches with retry loops, priority queuing, SQLite memory, pattern mining, scheduled tasks, and live TUI + web dashboard. Remote control via Telegram.
- 4
- Rust
- MIT
- 3 mo ago
opencrab
🦀 A self-owning territory crab — an autonomous digital life that owns its goals and runs a life loop. The next molt after OpenClaw and Hermes.
- 4
- Python
- MIT
- 6 mo ago
self-learning-24.7-agent
Expert Fleet - a stdlib-only, file-backed platform for 24/7 expert AI agents: frozen goal contracts, verified procedural learning (repeated gated work compiles into zero-model procedures), sealed capability packs, and a proof system where no status can be set by hand. 105 modules, 138 acceptance tests.
- 1
- Python
- AGPL-3.0
- 3 wk ago
factory-workbench
A local workbench that runs durable, reviewed jobs and safely upgrades its own orchestration: Lean 4 kernel on the admission path, sandboxed verifier, human approval.
- 0
- TypeScript
- MIT
- 8 wk ago
ARIA — Self-Improving AI
Self-improving AI that runs on 8GB RAM, no GPU needed
Runs in the browser on Hugging Face
- 0
- apache-2.0
- 8 mo ago
siyaq-context-engineering-paper
- 0
- 0
- mit
- 5 mo ago
HDCS
LLM loop that reasons in a dense technical register — mechanical gates + adversarial judge + reverse-translation for humans. Self-upgrading. Standalone or as a DSH plugin.
- 0
- JavaScript
- 2 mo ago



















