Hedwig: Dynamic Autonomy for Coding Agents Under Local Oversight

2026-05-12

AI Summary

Existing coding agents rely on static permission configurations, which fail to accommodate developers’ dynamic autonomy preferences across varying tasks and stages, leading to poor trust calibration and inefficient collaboration. This work proposes Hedwig, a command-line coding agent that introduces a novel dynamic autonomy mechanism grounded in the evolution of user trust. By continuously learning from developers’ feedback and decisions over multiple interaction rounds—and integrating conversation-level behavioral modeling with context-awareness—Hedwig adaptively adjusts its autonomy boundaries. Empirical results demonstrate that Hedwig minimizes unnecessary interventions during familiar tasks to enhance efficiency while increasing oversight in unfamiliar scenarios to ensure safety, thereby significantly reducing developers’ cognitive burden in trust calibration and validating the efficacy and feasibility of the dynamic autonomy paradigm.

Abstract

Despite coding agents' advances in handling increasingly complex tasks, their continued tendency to introduce unintended edits, subtle bugs, and scope drift that slip past code review means developers must still decide how much autonomy to grant them. However, existing approaches for setting an agent's level of autonomy, such as static permission settings or instruction files, cannot account for how developers' preferences for agent autonomy can shift across tasks and over time. We conducted a formative survey with 21 software engineers who use coding agents and found that they experience frustration with calibrating autonomy and have evolving preferences for level of oversight. Building on these insights, we present Hedwig, a CLI coding agent that dynamically adjusts its autonomy level based on developer-agent interactions across sessions. Rather than operating on a global, fixed autonomy configuration, Hedwig learns an evolving set of behavioral guidelines from developer decisions and feedback, reducing friction on work for which the agent has earned trust, while tightening oversight when the agent operates outside familiar territory. Hedwig demonstrates the potential of a new paradigm where agents intelligently adapt their level of autonomy based on user trust through active, longitudinal collaboration.

Problem

• dynamic autonomy
• coding agents
• developer oversight
• autonomy calibration
• trust adaptation

Innovation

• dynamic autonomy
• coding agents
• developer trust
• adaptive oversight
• behavioral guidelines