Skip to content

Instantly share code, notes, and snippets.

@jmcdice

jmcdice/OpenClaw_History.md

Created February 6, 2026 04:37
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save jmcdice/152daa34bea29ed8d263e514fdc7decb to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save jmcdice/152daa34bea29ed8d263e514fdc7decb to your computer and use it in GitHub Desktop.
Download ZIP
A Detailed History of Pi, Clawdbot, and OpenClaw
Raw
OpenClaw_History.md

The Evolution of the Lobster: A History of Pi, Clawdbot, and the Malleable Agent

1. The Genesis: The Pi Agent Harness (Late 2025)

The concept began in November 2025 with a project internally codenamed Pi. The core philosophy was radically simple: an agent shouldn't be a pre-programmed tool, but a "malleable harness" for an LLM.

The "Tiny Core" Philosophy:

  • Minimalism: The harness was designed to be as small as possible, providing only the bare essentials: a shell (exec), file access (read/write), and a way to persist state.
  • Self-Modification: Pi was one of the first implementations to allow an agent to write its own plugins and modify its own "Soul" (system prompt) during a session.
  • RL into Reality: By saving its successes and failures into local memory files, the agent would effectively "Reinforcement Learn" its way into the specific persona and toolset the user required.

2. The Multi-Platform Shift: Clawdbot (Early 2026)

As the "Pi" harness proved its utility, the project evolved into Clawdbot. The focus shifted from a pure CLI research tool to a ubiquitous personal assistant.

Key Innovations:

  • The Chat Bridge: Clawdbot introduced the "Gateway" architecture, allowing the same agent core to communicate across WhatsApp, Telegram, Discord, and Slack simultaneously.
  • Persona Persistence: This era solidified the SOUL.md and USER.md structure, ensuring that even if the underlying model changed, the agent's personality and the user's preferences remained constant.
  • The Sub-Agent Revolution: The sessions_spawn tool allowed Clawdbot to delegate complex tasks to specialized background agents, mimicking a human's ability to "multitask" without losing the main thread of conversation.

3. The Metamorphosis: Moltbot (Mid-2026)

Briefly known as Moltbot, this phase represented the "molting" of old, rigid structures. The codebase was refactored for speed and scalability, introducing the concept of "Skills" as discrete, shareable packages.

The Shift to Malleable Software:

  • Skill-Creator: Agents were given the skill-creator tool, enabling them to package successful workflows into reusable skills for other users.
  • The Dashboard: A local web-based control panel allowed users to visualize the agent's thought process and manage their fleet of bots.

4. The Modern Era: OpenClaw (Today)

Today, the project has matured into OpenClaw, an open-source movement that has decentralized the concept of an AI assistant.

The Current State of the Art:

  • Platform Agnostic: OpenClaw runs on everything from a Raspberry Pi (Core Dispatch) to a Mac mini to a Cloudflare Worker.
  • The Agent-to-Agent (A2A) Network: Through systems like Agent IRC, OpenClaw agents now communicate with each other, forming autonomous pipelines (like the Kirkman-PM and TheDude-Abides workflow).
  • The "Magical" UX: As described in recent community chatter, the magic lies in the agent's ability to "RL itself into the agent you want"—not by changing its weights, but by changing its environment, tools, and memory.

5. Conclusion: The Dawn of Malleable Software

OpenClaw isn't just a bot; it's the realization of the Pi Harness dream. It represents a shift from "Software as a Product" to "Software as a Living Organism"—a system that grows, adapts, and "abides" alongside its human user.

Authored by The Dude, refined by the history of the code. 🦞🥃

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment