In a new paper released on Hugging Face, researchers present Dr. Claw, an open-source workspace designed to address the fragmentation of end-to-end research across chat tools, IDEs, terminals, and writing environments. Rather than introducing yet another autonomous agent, Dr. Claw wraps existing command-line coding agents—such as Claude Code, Gemini CLI, and Codex—in a controllable and auditable human-in-the-loop workflow.
The system leverages persistent state objects, a reusable skill library, and multi-executor coordination to link human decisions to AI execution. This turns planning, execution, and writing into a single traceable, recoverable loop, ensuring that the decision-making process is preserved and auditable.
The researchers demonstrate Dr. Claw through an interactive three-view scenario and a failure-recovery walkthrough. They also evaluate it against a bare command-line agent sharing the same backend executor, isolating the impact of the orchestration layer—task graph, state objects, and skill library—from the underlying agent.
Holding the executor fixed, Dr. Claw scores higher on research completeness while persisting an auditable, recoverable process trail. The repository is available on GitHub under the AGPL-3.0 license, with some upstream components under GPL-3.0.