In , we introduced how Spotify uses a background coding agent to extend Spotify’s Fleet Management system. It can edit code, run builds and tests, and open pull requests (PRs) automatically. Once we had our agents up and running, a new challenge emerged:

The Spotify Engineering Blog provides  insights, technical deep dives, and behind-the-scenes stories from the engineering team at Spotify. Covering a wide range of topics, including backend systems, data processing pipelines, and frontend architecture, the blog offers a glimpse into the technology stack and engineering practices used to power one of the world's leading music streaming platforms. Developers can learn about scalable infrastructure solutions, innovative engineering techniques, and best practices for building high-performance software systems. By following the Spotify Engineering Blog, developers can gain inspiration, knowledge, and practical insights to apply to their own projects and endeavors.

Spotify Labs

Spotify shares lessons learned from implementing background coding agents for automated code migrations. The team evolved from open-source tools like Goose and Aider to building a custom agentic loop, eventually settling on Claude Code for handling complex, multi-file changes across thousands of repositories. Key insights include crafting effective prompts that describe desired end states rather than step-by-step instructions, stating preconditions clearly, using concrete examples, and limiting tool access to maintain predictability. The system successfully merged thousands of PRs by balancing prompt specificity with agent autonomy, though evaluation and feedback loops remain areas for improvement.

Background Coding Agents: Context Engineering (Part 2)