Researchers wrap coding agents in an auditable workflow, not a new agent

Dingjie Song and twelve co-authors, including Hanrong Zhang and Lichao Sun, present Dr. Claw, an open-source workspace that wraps command-line coding agents such as Claude Code and Gemini CLI in a human-in-the-loop workflow, rather than building a new autonomous agent. The system adds persistent state objects, a reusable skill library, and multi-executor coordination so that planning, execution, and writing form one traceable, recoverable loop instead of fragmenting across chat tools, IDEs, terminals, and separate writing environments. Tested against a bare command-line agent sharing the same backend executor, Dr. Claw scored higher on research completeness while preserving an auditable process trail. The paper, accepted to EMNLP 2026 System Demonstrations, releases the tool under AGPL-3.0 with GPL-3.0 upstream components.