AgentHazard benchmark shows computer-use agents remain highly vulnerable, with attack success rates reaching 73.63% on models like Qwen3-Coder powering Claude Code.
Locobench-agent: An interactive benchmark for LLM agents in long-context software engineering
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
citation-role summary
dataset 2
background 1
citation-polarity summary
years
2026 10representative citing papers
ECHO stores each agent turn as a source-indexed memory, reconstructs bounded contexts by selecting useful records, and routes RL credit through the same selection trace — reaching 43.4% on BrowseComp-Plus vs 28.9% (GRPO) and 36.1% (SUPO).
DeepDiscovery uses a two-stage framework to improve task-relevant file recovery and downstream performance in large codebases, achieving gains on industrial tasks and SWE-bench Verified.
RL Developer Memory is a feedback-normalized, safety-gated memory architecture for RL coding agents that logs contextual decisions and applies conservative off-policy gates to maintain 80% decision accuracy and full hard-negative suppression on a 200-case benchmark.
The paper defines a bounded reference architecture for LLM-orchestrated hybrid retrieval in dataset search using BM25, dense embeddings, reciprocal rank fusion, and metadata augmentation with pseudo-queries.
SWE-MeM introduces adaptive memory management for coding agents via synthesized trajectories and Memory-aware GRPO, reporting 43.4% and 60.2% resolve rates on SWE-Bench Verified for 4B and 30B models while beating baselines on performance and token use.
BraveGuard trains guard models on realistic agent trajectories derived from open-world threats, raising detection accuracy on AgentHazard from 38.79% to 82.38%.
Formal architecture descriptors reduce AI coding agent navigation steps by 33-44% and behavioral variance by 52% in controlled and observational studies.
UA-ChatDev integrates token-level uncertainty estimation and phase-aware verification into multi-agent software development and reports better benchmark scores than prior frameworks.
Survey framing LLM agents as model-plus-harness systems, decomposing harness responsibilities, mapping them to tasks, and highlighting open challenges in evaluation, safety, and co-evolution.
citing papers explorer
-
AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agents
AgentHazard benchmark shows computer-use agents remain highly vulnerable, with attack success rates reaching 73.63% on models like Qwen3-Coder powering Claude Code.
-
ECHO: Prune To Act, Trace To Learn With Selective Turn Memory In Agentic RL
ECHO stores each agent turn as a source-indexed memory, reconstructs bounded contexts by selecting useful records, and routes RL credit through the same selection trace — reaching 43.4% on BrowseComp-Plus vs 28.9% (GRPO) and 36.1% (SUPO).
-
From Fragments to Paths: Task-Level Context Recovery for Large Industrial Codebases
DeepDiscovery uses a two-stage framework to improve task-relevant file recovery and downstream performance in large codebases, achieving gains on industrial tasks and SWE-bench Verified.
-
Feedback-Normalized Developer Memory for Reinforcement-Learning Coding Agents: A Safety-Gated MCP Architecture
RL Developer Memory is a feedback-normalized, safety-gated memory architecture for RL coding agents that logs contextual decisions and applies conservative off-policy gates to maintain 80% decision accuracy and full hard-negative suppression on a 200-case benchmark.
-
A Reference Architecture for Agentic Hybrid Retrieval in Dataset Search
The paper defines a bounded reference architecture for LLM-orchestrated hybrid retrieval in dataset search using BM25, dense embeddings, reciprocal rank fusion, and metadata augmentation with pseudo-queries.
-
SWE-MeM: Learning Adaptive Memory Management for Long-Horizon Coding Agents
SWE-MeM introduces adaptive memory management for coding agents via synthesized trajectories and Memory-aware GRPO, reporting 43.4% and 60.2% resolve rates on SWE-Bench Verified for 4B and 30B models while beating baselines on performance and token use.
-
BraveGuard: From Open-World Threats to Safer Computer-Use Agents
BraveGuard trains guard models on realistic agent trajectories derived from open-world threats, raising detection accuracy on AgentHazard from 38.79% to 82.38%.
-
Formal Architecture Descriptors as Navigation Primitives for AI Coding Agents
Formal architecture descriptors reduce AI coding agent navigation steps by 33-44% and behavioral variance by 52% in controlled and observational studies.
-
UA-ChatDev: Uncertainty-Aware Multi-Agent Collaboration for Reliable Software Development
UA-ChatDev integrates token-level uncertainty estimation and phase-aware verification into multi-agent software development and reports better benchmark scores than prior frameworks.
-
From Question Answering to Task Completion: A Survey on Agent System and Harness Design
Survey framing LLM agents as model-plus-harness systems, decomposing harness responsibilities, mapping them to tasks, and highlighting open challenges in evaluation, safety, and co-evolution.