TRAJDEBUG: Tracing Error Lifecycle to Identify Critical Failures in Long-Horizon Agent Trajectories

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Yunjia Qi, Zehua Yin, Xintong Shi, Hao Peng, Songyuanyi Lu, Yixian Liu, Richeng Xuan, Yuhong Liu, Zhichao Hu, Xiaozhi Wang, Lei Hou, Bin Xu, Juanzi Li
• Venue: arXiv: cs.AI
• Published: 2026-08-06
• Collected: 2026-08-07
• URL: http://arxiv.org/abs/2608.06346v1
• PDF: https://arxiv.org/pdf/2608.06346v1
• DOI: UNKNOWN

Plain-Language Brief

The paper proposes a framework called TrajDebug to improve error detection in long-term AI agent trajectories. It addresses the challenges of identifying errors in lengthy trajectories and multiple local errors with varying effects by using multi-granularity history compression, evidence-based error identification, and tracing each error's resolution status and terminal impact.

What the Document Says

• The paper discusses the challenges of critical error detection in long-trajectory agentic systems due to scattered evidence and multiple local errors with different downstream effects.
• TrajDebug is proposed as a framework for addressing these challenges, using multi-granularity history compression, evidence-based error identification, and tracing each error's resolution status and terminal impact.
• 486 manually annotated failed trajectories from Tau2Bench and SWE-Bench Pro are used to construct TrajErrBench for testing the framework.
• Experiments across various agent benchmarks show that TrajDebug outperforms existing baselines.
• The paper claims that TrajDebug provides actionable feedback for improving downstream agent success.

Key Concepts Mentioned

• LLM-based agentic systems
• critical error detection
• long trajectories
• cascading errors
• evidence-based error identification
• error lifecycle tracing
• TrajDebug
• TrajErrBench
• Tau2Bench
• SWE-Bench Pro

Claims Made by the Document

• TrajDebug addresses the challenges of long-trajectory error discovery and critical attribution.
• TrajDebug outperforms existing baselines in various agent benchmarks.
• TrajDebug provides actionable feedback for improving downstream agent success.

Evidence Used

• The paper describes the proposed framework TrajDebug and its methodology.
• Experiments are conducted and results are presented to support the claims.

Methods / Approach

• multi-granularity history compression
• evidence-based error identification
• error lifecycle tracing

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

LLM-based agentic systems have shown remarkable capabilities in complex domains, while suffering from cascading errors and difficulty in debugging. Critical error detection aims to locate the earliest error step in a failed trajectory that is responsible for the final failure. However, progress faces two main challenges. First, long trajectories make it difficult to identify individual errors, since the evidence for judging a step may be scattered across distant instructions, observations, and prior context. Second, failed trajectories often contain multiple local errors with different downstream effects, only some of which remain responsible for the final failure. In this work, we propose TrajDebug, an error-lifecycle tracing framework that addresses long-trajectory error discovery with multi-granularity history compression and evidence-based error identification, and supports critical attribution by tracing each error's resolution status and terminal impact. We further construct TrajErrBench, a benchmark of 486 manually annotated failed trajectories from Tau2Bench and SWE-Bench Pro, covering realistic tool-use and coding scenarios. Experiments across diverse agent benchmarks show that TrajDebug achieves the best overall performance over existing baselines, and application studies further demonstrate that its diagnoses provide actionable feedback for improving downstream agent success. We will release the codes and data to facilitate further research.