NEWkern — a runtime for agent-kernel with Telegram, Slack, and daemon mode →

agent-kernel

Three markdown files that make any AI agent stateful.

Your agent remembers between sessions, takes notes, and builds on past work. No framework, no database — just three markdown files and a git repo.

$ git clone https://github.com/oguzbilgic/agent-kernel.git my-agent

$ opencode my-agent/

Works with any AI coding agent: OpenCode, Claude Code, Codex, Cursor, Windsurf, etc.

How it works

AI agents already read AGENTS.md as project instructions. The kernel uses that mechanism to teach the agent how to remember.

Clone the repo, start your coding agent, talk to it. It reads the kernel, realizes it's new, asks who you want it to be. You tell it. It remembers.

Memory structure

AGENTS.md          ← kernel (generic, don't edit)
IDENTITY.md        ← who this agent is
KNOWLEDGE.md       ← index of knowledge files
knowledge/         ← facts about the world (mutable)
notes/             ← daily session logs (append-only)

knowledge/

State. Facts about how things are right now. The agent updates these when reality changes. Each file tracks when it was last updated.

notes/

Narrative. What happened each session — decisions, actions, open items. Append-only. Never modified after the day ends.

Quick start

Clone the kernel

git clone https://github.com/oguzbilgic/agent-kernel.git my-agent

Start your agent

cd my-agent && opencode

Talk to it

Tell it who it is and what it should do. It remembers.

Works with opencode, claude, codex, cursor, or any coding agent that reads AGENTS.md.

Multiple agents

Each agent is its own repo. Same kernel, different identity, different knowledge.

git clone agent-kernel homelab-agent
git clone agent-kernel investing-agent
git clone agent-kernel health-agent

Want Telegram, Slack, and a daemon?

kern-aiis a runtime built for agent-kernel. One brain across every channel — TUI, Telegram, Slack. Daemon mode, user pairing, backup & restore.

$ npx kern-ai init my-agent