AutoGPT Introduces New AI-Contributor Management System to Tackle Rising Pull Request Volume

AutoGPT, the open‑source AI project that has attracted over 180,000 stars, is grappling with a surge of pull requests generated by tools such as Copilot, OpenClaw and its own internal agents. The project’s maintainers, led by founding AI engineer Nicholas Tindle, have implemented a suite of safeguards to keep the review process efficient and human‑friendly. Rather than shutting down AI contributions outright, AutoGPT has placed specialized instruction files—AGENTS. md and CLAUDE. md—directly beside the code they affect, ensuring that agents receive the right context. Pull requests must now follow a strict template that includes a test plan, and automated checks enforce coverage thresholds and a signed Contributor License Agreement (CLA). If a pull request fails a check, the system will close it and prompt the contributor to sign the CLA before reopening. The team also introduced a “pr‑address” skill that requires a commit SHA before a review thread can be resolved, preventing bots from marking issues as fixed without code changes. While the project once had bots comment on every CI failure, that feature was disabled to reduce noise. AutoGPT’s approach highlights the growing need for clear guidelines and automated gates when AI tools become a primary source of code contributions.