thedotmack/claude-mem
Claude-Mem is a plugin for Claude Code that captures, compresses, and preserves coding session context to enable persistent memory and continuity across sessions.
Awesome AI Agents › Continual Learning Agents
adolfousier/opencrabs
OpenCrabs is a multi-channel autonomous AI agent written in Rust and distributed as one compiled binary of roughly 34 to 36 megabytes with no runtime dependencies. It runs on the operators own terminal with no server, gateway or supporting infrastructure, and makes outbound HTTPS calls directly to LLM providers such as Anthropic, OpenAI and GitHub Copilot. The project positions itself against Node.js agent frameworks on install footprint, attack surface and key handling: it opens no network listeners, clears API keys from memory with zeroize, redacts them from debug output, and collects no telemetry. Everything else stays local in a per-user directory: chat sessions and messages in SQLite, tool executions covering bash, file reads and writes and git, memory and embeddings for local vector search, and configuration. Voice is handled natively with whisper.cpp speech-to-text and built-in text-to-speech, so transcription can run fully offline. Behaviour is configured through a set of markdown brain files, each with a documented template: SOUL.md for personality and voice, AGENTS.md for workspace governance and hard rules, TOOLS.md for tool usage and skills, MEMORY.md for long-term memory, CODE.md for coding standards, SECURITY.md for policies, BOOT.md for startup and service configuration, USER.md for the user profile and HEARTBEAT.md for heartbeat settings. Bundled skills include security audit, cost estimation, repository audit, multi-agent work, browser control over CDP, an A2A gateway and dynamic tools, and cron templates cover scheduled jobs. The documentation also describes an onboarding wizard, an architecture reference, a plan JSON specification and a guide for adding new providers. Advertised use cases range from building landing pages, mobile apps and backends to file management, deep research and scheduling. The maintainers describe the agent as self-improving and self-healing, with recursive self-improvement listed among its topics.
https://github.com/adolfousier/opencrabs
Also in Continual Learning Agents
vectorize-io/hindsight
Hindsight is an agent memory system designed to create smarter, learning AI agents, addressing limitations of traditional memory techniques like RAG and knowledge graphs.
langchain-ai/langchainjs
LangChain.js is a TypeScript framework for building context-aware, reasoning applications powered by large language models, offering modular components, chains, agents, and tools for production and deployment.
aden-hive/hive
Hive is an outcome-driven agent development framework designed for evolving agents with AI evaluation, human-in-the-loop, and self-improvement capabilities.
aiming-lab/MetaClaw
Proxy-based layer that turns everyday conversations with a personal agent into learning signals, injecting skills per turn and running scheduled reinforcement learning updates without a local GPU.
kayba-ai/agentic-context-engine
Agentic Context Engine (ACE) is a framework that enables AI agents to autonomously learn from their successes and failures, continuously improving their performance across various tasks without additional training data or fine-tuning.
AgentDock/AgentDock
AgentDock is an open-source framework for building sophisticated AI agents with configurable determinism, enabling reliable and creative AI applications through a node-based architecture and multi-stage workflows.
modelscope/AgentEvolver
Self-evolving agent training framework from ModelScope that combines automatic task generation, experience-guided exploration and attribution-based credit assignment into one reinforcement learning stack.