HarnessNet: Feedback-driven Hierarchical Joint Harness Evolution for Multi-agent Systems
cs.AI
Submitted: 2026-10-10
Updated: 2026-10-10
Code: https://github.com/ollama/ol
Terminology
Sources
- Live-SWE-agent: Can Software Engineering Agents Self-Evolve on the Fly?
- Meta-Harness: End-to-End Optimization of Model Harnesses
- Adapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM Agents
- Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses
- Continual Harness: Online Adaptation for Self-Improving Foundation Agents
- Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams
- HarnessEvolve: Learning from Reference Trajectories for Reliable Agent Self-Evolution
- AgentLeak: A Benchmark for Internal-Channel Privacy Leakage in Multi-Agent LLM Systems
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
- Generalizable Self-Evolving Memory for Automatic Prompt Optimization
- Context Language Models
- Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models
- Dynamic Cheatsheet: Test-Time Learning with Adaptive Memory
- AutoSkill: Experience-Driven Lifelong Learning via Skill Self-Evolution
- MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation
- CoEvoSkills: Self-Evolving Agent Skills via Co-Evolutionary Verification
- SkillGrad: Optimizing Agent Skills Like Gradient Descent
- MetaSkill-Evolve: Recursive Self-Improvement of LLM Agents via Two-Timescale Meta-Skill Evolution
- A Self-Improving Coding Agent
- Darwin Godel Machine: Open-Ended Evolution of Self-Improving 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
- Agents' Last Exam
- 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