thepopebot  by stephengpope

Autonomous AI agent for continuous, auditable task execution

Created 3 weeks ago

New!

742 stars

Top 46.7% on SourcePulse

GitHubView on GitHub
Project Summary

Thepopebot provides a framework for creating autonomous AI agents where the Git repository itself serves as the agent's state and action log. This approach offers unparalleled auditability and reversibility, as every agent action is a Git commit. It targets developers and power users seeking transparent, controllable AI agents, leveraging free GitHub Actions compute for execution and enabling self-evolution through auditable pull requests.

How It Works

The system operates via a Git-centric workflow. An event handler initiates jobs, which are translated into Git branches. GitHub Actions then provisions a Docker container running a "Pi coding agent" to execute tasks. The agent commits its work and opens a pull request, which can be auto-merged or trigger further workflows. This design ensures all agent activities are version-controlled, auditable, and easily reversible, with the repository acting as the definitive record.

Quick Start & Requirements

Setup requires Node.js 18+ (nodejs.org), npm, Git (git-scm.com), GitHub CLI (cli.github.com), Docker (docker.com) with Docker Compose, and ngrok (ngrok.com) for local development webhook access. Installation involves scaffolding with npx thepopebot@latest init, followed by an interactive setup wizard (npm run setup) that configures the GitHub repository, API keys (Anthropic required; OpenAI, Brave optional), and builds the project. The agent is then launched via docker compose up -d.

Highlighted Details

  • Git as Agent State: Every action is a Git commit, enabling full history, rollback, and easy forking of agent configurations and personalities.
  • Free Compute: Leverages included GitHub Actions compute for running agent jobs and managing pull requests, scaling from one to hundreds of parallel tasks.
  • Self-Evolving Code: Agents can modify their own code via pull requests, with all changes being auditable and reversible.
  • Transparent Operations: Provides a clear, auditable trail of agent actions, enhancing trust and control.

Maintenance & Community

The provided README does not detail specific maintainers, community channels (like Discord or Slack), or sponsorship information.

Licensing & Compatibility

The README does not specify the project's license or provide compatibility notes for commercial use.

Limitations & Caveats

Local development setups using tunnels expose public endpoints without robust security measures like rate limiting or TLS on the local hop. Users are responsible for securing their own infrastructure. The system requires specific API keys, and while it offers automated upgrades, recovery from failed upgrades is also documented.

Health Check
Last Commit

19 hours ago

Responsiveness

Inactive

Pull Requests (30d)
126
Issues (30d)
16
Star History
753 stars in the last 23 days

Explore Similar Projects

Starred by Eric Zhu Eric Zhu(Coauthor of AutoGen; Research Scientist at Microsoft Research).

gh-aw by github

15.6%
4k
Agentic workflows for GitHub automation
Created 6 months ago
Updated 16 hours ago
Starred by Lilian Weng Lilian Weng(Cofounder of Thinking Machines Lab), Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems"), and
59 more.

AutoGPT by Significant-Gravitas

0.1%
182k
AI agent platform for building, deploying, and running autonomous workflows
Created 2 years ago
Updated 15 hours ago
Feedback? Help us improve.