Preprint
Jul 2026
When and How Context Rot Appears in Coding Agents: A White-Box Study of Agent Skills in Code Auditing
This work provides a bounded failure classification and empirical case study for white-box code auditing, and varies the surrounding context and classify where failures first become visible: lost requirements, editing drift, failed checking, or non-agent evaluator/runtime failures.
Yue Xue
· 0 citations