Collaborator or Assistnat? How AI Coding Agents Partition Work Across Pull Request Lifecycles
Young Jo, Chung, Safwat Hassan
TLDR
This paper analyzes how AI coding agents partition work in pull request lifecycles, classifying them as Collaborators or Assistants based on their agency.
Key contributions
- Characterizes AI coding agents on a Collaborator-Assistant spectrum for PR work partitioning.
- Analyzes 29,585 PR lifecycles, identifying six interaction scenarios using an Initiator x Approver taxonomy.
- Finds Collaborator tools (e.g., Devin) agent-initiate most PRs, while human merge authority remains dominant.
- Contributes a taxonomy, per-tool state machines, and a replication package for further research.
Why it matters
This research provides a crucial framework for understanding the evolving roles of AI in software development workflows. By distinguishing between agent initiative and human oversight, it highlights critical considerations for designing future AI-powered coding tools and governance mechanisms.
Original Abstract
When AI coding agents open branches and submit pull requests (PRs), two questions co-determine oversight design: who starts the work (operational agency) and who authorizes its completion (merge governance). We characterize tools along a Collaborator-Assistant spectrum in how they redistribute initiative, oversight, and endorsement, while merge governance remains predominantly human across five tools (OpenAI, Copilot, Devin, Cursor, Claude Code). We analyze 29,585 PR lifecycles using an Initiator x Approver taxonomy with six interaction scenarios; lifecycle reconstruction supplies the how behind those roles. Collaborator tools (Cursor, Devin, Copilot) concentrate operational initiative in agents that open and carry PR work forward, with humans retaining review and endorsement on the path to merge; Assistant tools (OpenAI, Claude) leave task direction primarily with humans and supply bounded support within human-led workflows. Across the spectrum, agency and governance decouple: Collaborator workflows are >=96% agent initiated, yet terminal merge authority remains almost exclusively human, with agent-classified approvers confined to a small fraction of PRs. Where automation executes a merge, logs record the executor but not the decision-maker, marking a boundary of observation. We contribute the taxonomy, per-tool state machines, and a replication package for research on automation, oversight, and governance in PR workflows.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.