A new arXiv method, FC-SWE, uses failure-conditioned trajectories in reinforcement learning to help long-horizon software engineering agents recover from failed patches by reusing diagnostic feedback from…

#SWEAgents #ReinforcementLearning #DevTools #AIResearch
https://arxiv.org/abs/2610.07898

New arXiv work shows 10.7% of passing SWE-agent trajectories are "Lucky Passes" with regressions or blind retries, highlighting that binary pass/fail signals hide messy agent behavior. AgentLens offers…

#SWEAgents #AgentEval #DeveloperTools #AIEngineering
https://arxiv.org/abs/2605.12925

A new token- and execution-free filter called LatentSift aims to make test-time verification cheaper for SWE agents by reusing the policy's own hidden states instead of running another LLM pass.

#SWEagents #LLM #AIresearch #devtools
https://arxiv.org/abs/2609.36371

SWE-MILE introduces an asynchronous potential-based framework for assigning milestone-level credit in long-horizon SWE agents, deriving process supervision from workflow runtime without extra reward models. It…

#SWEAgents #ReinforcementLearning #SoftwareEngineering
https://arxiv.org/abs/2609.32631