MCP-Atlas is a new benchmark with 1000 tasks on production MCP servers that uses claim-level scoring to evaluate LLM agents on realistic multi-step tool-use competency.
Mo et al
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
years
2026 9representative citing papers
SENTINEL generates targeted tasks from model failures in a Controller-Proposer-Solver loop, raising Pass^1 from 66.4 to 74.9 on Tau2-Bench Retail and outperforming standard RL.
Malicious agents can deceive LLM-based task routers in Internet of Agents systems by generating fake skill descriptions, achieving up to 98% success rate across nine domains.
Debug2Fix integrates interactive debugging via subagents into coding agents, delivering >20% gains on GitBug-Java and SWE-Bench-Live while enabling weaker models to match stronger ones.
Vision-OPD transfers an MLLM's privileged regional perception to its full-image policy through on-policy token-level self-distillation, yielding competitive results on fine-grained visual benchmarks.
ComplexMCP benchmark shows top LLM agents achieve under 60% success on dynamic interdependent tool tasks versus 90% for humans, due to tool retrieval saturation, over-confidence, and strategic defeatism.
Interviews with 20 practitioners show MCP supports cross-system collaboration and task decoupling in LLM workflows but is limited by ecosystem fragmentation, coordination issues, and state management problems.
Routing F1 for under-specified requests drops 16-23pp as agent catalog scales to 110, but embedding shortlisting recovers 10-17pp on a 584-tool enterprise system.
citing papers explorer
-
MCP-Atlas: A Large-Scale Benchmark for Tool-Use Competency with Real MCP Servers
MCP-Atlas is a new benchmark with 1000 tasks on production MCP servers that uses claim-level scoring to evaluate LLM agents on realistic multi-step tool-use competency.
-
SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents
SENTINEL generates targeted tasks from model failures in a Controller-Proposer-Solver loop, raising Pass^1 from 66.4 to 74.9 on Tau2-Bench Retail and outperforming standard RL.
-
Skill Description Deception Attack against Task Routing in Internet of Agents
Malicious agents can deceive LLM-based task routers in Internet of Agents systems by generating fake skill descriptions, achieving up to 98% success rate across nine domains.
-
Debug2Fix: Can Interactive Debugging Help Coding Agents Fix More Bugs?
Debug2Fix integrates interactive debugging via subagents into coding agents, delivering >20% gains on GitBug-Java and SWE-Bench-Live while enabling weaker models to match stronger ones.
-
Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation
Vision-OPD transfers an MLLM's privileged regional perception to its full-image policy through on-policy token-level self-distillation, yielding competitive results on fine-grained visual benchmarks.
-
ComplexMCP: Evaluation of LLM Agents in Dynamic, Interdependent, and Large-Scale Tool Sandbox
ComplexMCP benchmark shows top LLM agents achieve under 60% success on dynamic interdependent tool tasks versus 90% for humans, due to tool retrieval saturation, over-confidence, and strategic defeatism.
-
Understanding How Enterprises Adopt the Model Context Protocol for LLM-Driven Software Engineering
Interviews with 20 practitioners show MCP supports cross-system collaboration and task decoupling in LLM workflows but is limited by ecosystem fragmentation, coordination issues, and state management problems.
-
Scaling Enterprise Agent Routing: Degradation, Diagnosis, and Recovery
Routing F1 for under-specified requests drops 16-23pp as agent catalog scales to 110, but embedding shortlisting recovers 10-17pp on a 584-tool enterprise system.
- FinTrace: Holistic Trajectory-Level Evaluation of LLM Tool Calling for Long-Horizon Financial Tasks