Wink: Recovering from Misbehaviors in Coding Agents

2026-02-18

AI Summary

This work addresses the frequent disruptions in autonomous coding agents caused by instruction deviation, repetitive loops, or erroneous tool usage, which often necessitate extensive human intervention. To mitigate this, the authors propose Wink, a lightweight asynchronous self-intervention system that, for the first time, identifies three major categories of agent misuse based on real-world production traffic. Wink introduces a scalable automatic intervention mechanism that corrects agent behavior without halting execution, leveraging large language model–based trajectory monitoring, asynchronous intervention strategies, and real-time feedback. Evaluated on over 10,000 real agent trajectories, Wink successfully rectifies 90% of misuse instances with a single intervention. A/B testing further demonstrates significant reductions in tool invocation failure rates, per-session token consumption, and the frequency of manual interventions.

Technology Category

• Multiagent Systems: Agent Communication
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Autonomous coding agents, powered by large language models (LLMs), are increasingly being adopted in the software industry to automate complex engineering tasks. However, these agents are prone to a wide range of misbehaviors, such as deviating from the user's instructions, getting stuck in repetitive loops, or failing to use tools correctly. These failures disrupt the development workflow and often require resource-intensive manual intervention. In this paper, we present a system for automatically recovering from agentic misbehaviors at scale. We first introduce a taxonomy of misbehaviors grounded in an analysis of production traffic, identifying three primary categories: Specification Drift, Reasoning Problems, and Tool Call Failures, which we find occur in about 30% of all agent trajectories. To address these issues, we developed a lightweight, asynchronous self-intervention system named Wink. Wink observes agent trajectories and provides targeted course-correction guidance to nudge the agent back to a productive path. We evaluated our system on over 10,000 real world agent trajectories and found that it successfully resolves 90% of the misbehaviors that require a single intervention. Furthermore, a live A/B test in our production environment demonstrated that our system leads to a statistically significant reduction in Tool Call Failures, Tokens per Session and Engineer Interventions per Session. We present our experience designing and deploying this system, offering insights into the challenges of building resilient agentic systems at scale.

Problem

• coding agents
• misbehaviors
• large language models
• autonomous software development
• agent failures

Innovation

• agentic misbehavior
• self-intervention
• LLM-based coding agents
• tool call failures
• specification drift