Atlas

Atlas is my self-hosted personal operating agent built on OpenClaw. Instead of treating an AI system as a generic chatbot, the project defines it as a bounded operator with an explicit identity, clear tool permissions, a durable memory model, and a narrow set of approved surfaces for work.

At the center is a modular governance pack (AGENTS.md, BOOTSTRAP.md, HEARTBEAT.md, IDENTITY.md, SOUL.md, TOOLS.md, USER.md) that controls startup, policy enforcement, operating cadence, and change management. Local memory lives in MEMORY.md, Notion acts as a second brain for structured notes and execution artifacts, Readwise supports reading workflows, a runtime-configured web search path handles live retrieval, and Agentmail enables tightly constrained email handling through a trusted allowlist.

The system is designed to stay natural at the interface while being much more disciplined underneath. Atlas can move between conversation, recall, research, synthesis, drafting, and delegated execution, but it does so inside an explicit governance model rather than through vague prompt instructions or unchecked autonomy. Internally, the system uses one main conversational agent plus specialist roles for research, writing, and operator-style execution, so it can delegate work without turning into a noisy multi-bot interface.

One example of that operating model is DIGEST, a merged workflow that turns recent Readwise saves or newsletter material into clear, plain-English briefings, writes them to Notion, emails them out, and archives the Readwise items it has already summarized.

The goal is practical leverage without black-box behavior: better recall, better synthesis, and more useful delegation, while keeping the system local-first, auditable, and intentionally constrained.