Type: url
Original: https://github.blog/ai-and-ml/github-copilot/agent-driven-development-in-copilot-applied-science/
Status: generated
Summary
A researcher on GitHub’s Copilot Applied Science team describes using Copilot CLI and the Copilot SDK to build eval-agents, a set of agents that automate the analysis of coding-agent benchmark trajectories, with coding agents as the project’s primary contributor. The post distils three principles — conversational, plan-first prompting; frequent refactoring, documentation and cleanup; and a “blame process, not agents” approach backed by tests, strict typing and linters — into a development loop of planning, autopilot implementation, an automated code-review loop and human review.