Do Not Restart: Residual Completion for Stateful Agent Handoffs

arXiv:2609.13800 · cs.AI · Submitted 2026-09-12 · Read on arXiv

cs.AI

Submitted: 2026-09-12

Updated: 2026-09-26

Comments: 11 pages, 2 figures, 4 tables

Code: https://github.com/microsoft/STATE-Bench

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Checked Execution — 8× in this paper
• Benchmark — 7× in this paper · explained in 2 episodes
• Accuracy — 6× in this paper · explained in 1 episode
• Grounding — 6× in this paper · explained in 3 episodes
• Commitment-Frontier Residual Completion — 3× in this paper
• Fidelity — 3× in this paper · explained in 2 episodes
• Benchmarking — 2× in this paper · explained in 3 episodes
• Reached-State Contract Construction — 2× in this paper
• Verification — 2× in this paper · explained in 2 episodes
• AI Agents — 1× in this paper · explained in 2 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Core — 1× in this paper · explained in 2 episodes
• Embeddings — 1× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Provenance — 1× in this paper · explained in 3 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Trace — 1× in this paper · explained in 2 episodes

Sources

• From Confident Closing to Silent Failure: Characterizing False Success in LLM Agents
• $\tau^2$-Bench: Evaluating Conversational Agents in a Dual-Control Environment
• Is Escalation Worth It? On the Depth of LLM Cascades
• FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance
• Cordon: Semantic Transactions for Tool-Using LLM Agents
• Defeating Prompt Injections by Design
• The Granularity Mismatch in Agent Security: Argument-Level Provenance Solves Enforcement and Isolates the LLM Reasoning Bottleneck
• Can Agent Benchmarks Support Their Scores? Evidence-Supported Bounds for Interactive-Agent Evaluation
• Enforcing Temporal Constraints for LLM Agents
• Handoff Debt: The Rediscovery Cost When Coding Agents Take Over Interrupted Tasks
• VIGIL: Defending LLM Agents Against Tool Stream Injection via Verify-Before-Commit
• Agentic Routing: The Harness-Native Data Flywheel
• ToolGate: Contract-Grounded and Verified Tool Execution for LLMs
• ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilities
• RouteLLM: Learning to Route LLMs with Preference Data
• GoEX: Perspectives and Designs Towards a Runtime for Autonomous LLM Applications
• Agent-Diff: Benchmarking LLM Agents on Enterprise API Tasks via Code Execution with State-Diff-Based Evaluation
• SWE-Router: Routing in Multi-turn Agentic Software Engineering Tasks
• Agent Workflow Memory
• Step-level Optimization for Efficient Computer-use Agents

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems
• VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection