Video · GitHub ·
Four days, zero lines: the AI force multiplier with GitHub Copilot
In short
Jeff Doolittle explains how he used GitHub Copilot and GitHub Actions to build an OSS governance system in four days without personally writing code. The approach turned a 436-repository compliance backlog into a reversible, self-service workflow, showing that AI amplifies human architectural judgment rather than replacing it.
Key points
- The backlog covered 436 of 439 repositories, leaving only 3 compliant.
- Copilot generated deterministic Python automation using GitHub’s GraphQL API, parallel execution, rate-limit handling, and state persistence.
- Human judgment defined exceptions for forks, code owners, test-tool vulnerabilities, and policy enforcement.
- The system issues remediation tickets, warns on day 21, grants automatic 14-day extensions, and enforces reversible visibility changes on day 30.
- Self-service Copilot remediation helped 30 repositories resolve violations within one week.
- Public repositories reached 100% compliance, while 12 volunteers joined to maintain the system.
- The recommended pattern is to specify intent and constraints, encode expert knowledge, and reduce friction for audits, onboarding, policies, and technical debt.
Written by an AI model from the video’s captions and description.
How do you clear a massive governance backlog without destroying developer velocity? In just four days, Jeff Doolittle of Trimble used GitHub Copilot to build an automated system and rapidly clear the backlog without manually writing a single line of code.
In this story
A tool or model opens in the Heidelberg AI Curriculum: what it does, what it costs, how to start. A repo opens on GitHub.


