Agent 系统、运行时与自我改进
知识导航
主题综述
概念与方法
- Harness Engineering
- Context Engineering
- Self-Improving Harness
- Recursive Self-Improvement
- Meta-Agent
- Reversible Agentic Execution Trace
- 内容选题
- Agent Plugin
- Harness Optimization
- Agent Failure Localization
- Agentic RL
人物、组织、产品与工具
- Lilian Weng
- Shepherd
- Agent Plugins
- HarnessOpt-Bench
- Interaction-Centric Agent Failure Taxonomy
- Agentic ESOpt
- Recuris
- SkillForge
- StarHarness
- HarnessLens
- PILOT
- Astar
- Meta-Harness
- Self-Harness
- Agentic Harness Engineering
来源摘要
- Harness Engineering for Self-Improvement
- Shepherd: Enabling Programmable Meta-Agents via Reversible Agentic Execution Traces
- 14年老编辑真诚分享：十大通用选题模板
- Agent Plugins
- HarnessOpt-Bench: Evaluating LLMs at Harness Optimization
- Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent Failures
- Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Memory Requirements
- FACET: Preserving Source Intent and Executable State in Terminal Task Synthesis
- Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses
- SkillForge: Evolving Verifiable Skills for Reinforcement Learning Agents
- StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments
- Verify Smarter, Evolve Further: Efficient Harness Evolution through Behavior-Aware Verification
- PILOT in the Loop: Live Self-Improvement for Long-Horizon Agents
- Astar: Learning to Propose Evolution Directions for Self-Evolving Industrial AI Systems
- Meta-Harness: End-to-End Optimization of Model Harnesses
- Self-Harness: Harnesses That Improve Themselves
- Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses