When May an Agent Stop? Evidence-Carrying Termination for Tool-Using LLMs
Source
arxiv.org
Author
Jason Liu
Date
Why it matters
Letting a model judge its own doneness breaks under lost responses and partial commits; requiring a replayable evidence certificate before COMPLETE is a concrete gate you can put in your own agentic loopThe cycle an agent runs in: decide, call a tool, read the result, decide again — repeating until the goal is met or a stop condition fires.Full definition →.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
agentic loop — The cycle an agent runs in: decide, call a tool, read the result, decide again — repeating until the goal is met or a stop condition fires.
articleAssurance Envelopes for Autonomous Coding Agents: Minimum-Cost Evidence for Software ChangeAnjan Goswami
articleVerification as an Architectural Layer for LLM Agents: A V-Model Design, and a Pilot Study of Its Deterministic CoreAli Afoud, Jie JW Wu
articleGraphEcho: Structural Redundancy and Evidence Provenance in LLM Graph AgentsSikun Wang, Yixi Zhou, Lei Fan, Fan Zhang