The problem it solves

AI-Powered Onchain Project Manager

An AI-powered onchain project manager that handles the software development flow—from idea to delivery—without traditional human coordination.

Workflow

1. Client Onboarding

As a client, you start by verifying your identity and explaining your project to an AI agent. The agent understands your needs and creates a GitHub repository, setting the stage for development.

2. Developer Bidding

It then opens the project for bidding, allowing developers to apply. Each developer's GitHub is analyzed by an AI agent, which scores their experience. Their proof is verified onchain and they are onboarded onto the project.

3. Developer Assignment

Once selected, the developer gets added as a collaborator and receives a dedicated branch to work on.

4. Development & Submission

After completing their module, the developer submits the module for review. The AI agent tests the feature in an AI-automated browser.

5. Review & Merge

• If everything works as expected, it releases funds to the developer, opens a pull request, and merges the work to the main branch.
• If not, it raises a GitHub issue and provides feedback.

Features

• AI agents handle project management, quality control, and fund distribution.
• All transactions and interactions are recorded onchain in the Base network.
• The goal is to make project execution smooth and efficient, powered by onchain AI agents and zk proofs.

Challenges we ran into

Core Technology

The project’s core is the Self onchain proof system, which uses zero-knowledge proofs for secure identity verification. This allows clients and developers to onboard privately without revealing sensitive data.

We also integrated GitHub verification onchain to assess and score developers based on their GitHub profiles, ensuring only qualified individuals are selected for the project.

Custom Plugins with Goat SDK

A key component of the project was the development of two custom plugins for Vercel and GitHub within the Goat SDK. These plugins enabled automation of:

• GitHub repository interactions
• Deployment workflows on Vercel
• Repository creation
• Code pushes
• Pull requests

These integrations bridged the AI agents with the development workflow seamlessly.

Onchain Infrastructure

We deployed the project’s smart contracts on the Base network, ensuring that all transactions and fund releases were:

• Secure
• Scalable
• Eco-friendly

Automated Testing & Fund Release

For module testing, we used the Anchor AI browser to automatically test code in a controlled AI environment before any fund release. This ensured that everything worked as expected prior to approving contributions.

Highlights

A notable achievement was the creation of custom plugins that enabled:

• Seamless service integration
• Fully automated testing
• Smart contract-based fund releases
• End-to-end AI-powered project management