The paper proposes the Communicating Finite Automata (CFA) abstraction and a framework for component program refactoring in large language model inference services, achieving significant speedup and r…

The paper argues that web agents should abandon the reactive ReAct paradigm in favor of a plan-then-execute approach, which requires developing typed, task-level APIs to properly structure web interac…

This paper empirically demonstrates that the choice of plan representation (e.g., checklist vs. narrative) significantly impacts the robustness and success rate of LLM-based web agents.

This paper presents SpecBox, a runtime system for LLM agents that uses speculative sandbox preallocation to improve resource utilization and reduce interactive tail latency.

The paper identifies that the convenience of host-acting agents leads to semantic under-specification in user goals, which forces the agent to generate potentially risky execution plans.

PlanGuard is a training-free defense framework that uses an isolated Planner and hierarchical verification to defend LLM agents against Indirect Prompt Injection by verifying the consistency of planne…

The paper introduces Semantic Intent Fragmentation (SIF), an attack class demonstrating that multi-agent AI orchestrators can violate security policies through a composition of individually benign sub…

This paper introduces IntentTune, a framework for inferring user intent from under-specified queries in e-commerce search using user-specific behavioral signals and population-level demand patterns.

This paper introduces the Goal-Oriented Dialogue Runtime (GODR), a framework-neutral design pattern for managing complex, multi-domain, interruptible conversations with multiple interdependent objecti…

The paper presents a tool-making pipeline for production LLM agents that compiles repeated steps into validated, versioned tools before deployment, reducing latency and error rate.

The paper introduces new benchmarks for complex asynchronous planning and demonstrates that general constraint satisfaction formalizers (like CP-SAT) significantly outperform direct LLM planning or tr…

This tutorial explores advances and challenges in deploying large language model-based agentic systems across industries, with a focus on reasoning and planning, multi-agent coordination, and evaluati…

This paper documents and analyzes the failure process of strategies used to address conceptual drift in long-horizon LLM collaboration and introduces the concept of 'Index Sickness' and the 'Pang Prin…

The paper proposes defining 'intent-to-execution integrity' as the necessary end-to-end correctness property for securing LLM agents, arguing that current defenses are insufficient due to untrusted co…

SkillsInjector proposes a two-stage adaptive method to dynamically optimize skill selection, quantity, and presentation for LLM agents, significantly improving task performance over static injection m…

This paper investigates indirect prompt injection vulnerabilities in ReAct agents by systematically analyzing how the injection depth and payload framing affect attack success rates, finding that inje…

The paper investigates indirect prompt injection vulnerabilities in ReAct agents by systematically varying the injection depth, payload framing, and turn budget, finding that injection depth is the do…

The paper introduces SMH-Bench, a comprehensive benchmark built on a simulator to rigorously test LLM agents' ability to perform complex, environment-grounded reasoning and actions in realistic smart-…

This paper introduces PDAGENT-BENCH, a comprehensive benchmark for evaluating Large Language Models and vision-language models in the physical design stack of Very Large-Scale Integrated Circuits.

The paper introduces an LLM-agent framework to solve the 'last-mile forecasting' problem, bridging the gap between raw statistical predictions and business-ready forecasts by incorporating weakly stru…