← All courses and workshops

Course and workshop

Trunk-based development with collaborative agents

Try trunk-based development with small changes, fast feedback and working agreements that make human–AI collaboration safe.

Audience
Software teams and technical leaders looking to improve flow, feedback and collaboration.
Format
Hands-on workshop, arranged with you.
Duration
Tailored to your goals and format.

This hands-on workshop treats trunk-based development as an end-to-end way of working: changes stay small, integration happens often, feedback comes quickly and main remains ready to release. We also explore how AI agents can contribute without blurring responsibility for quality or decisions.

Trunk-based practice

  • Small tasks, short-lived branches and frequent integration to trunk.
  • A clean, linear and releasable main, with each atomic change traceable to a specific piece of work.
  • Automated checks, tests and builds as fast feedback, bringing quality assurance close to the work instead of saving it for a late gate.
  • Tests focused on observable behavior rather than implementation details, making change and refactoring easier.
  • Issue-centred work logs that keep decisions, questions and progress with the task.

Working with agents

  • Give agents a clear specification, acceptance criteria and relevant repository instructions.
  • Make the repository a predictable workspace with containers, stable test data and checks that can also run locally.
  • Let agents work on short-lived branches and propose changes; people remain accountable for decisions, review and integration to main.
  • Use test and check results to give concrete feedback to both the agent and the team. When collaboration exposes ambiguity, improve the specification, repository instructions or guardrails.
  • Choose review practices that fit your workflow: unnecessary hand-offs and manual approval gates can add waiting time, while focused review and clear ownership still matter.

What you will take away

Try a non-blocking workflow on a real task and identify what your specifications, repository, tests and team practices need in order to integrate frequently without compromising quality or accountability.

Further reading