Trajectory-Guided Fault Localization for Agent Skill Evolution
cs.SE, cs.AI
Submitted: 2026-10-08
Updated: 2026-10-08
Comments: 20 pages, 3 figures
Code: https://github.com/skillberry-ai/cap-evolve
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- EvoSkill: Automated Skill Discovery for Multi-Agent Systems
- Evaluating Large Language Models Trained on Code
- DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
- SkillAxe: Sharpening LLM-Authored Agent Skills Through Evaluation-Guided Self-Refinement
- SWE-Skills-Bench: Do Agent Skills Actually Help in Real-World Software Engineering?
- SkillAdam: Stable and Efficient Skill Evolution for Agents
- CodeTracer: Towards Traceable Agent States
- SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks
- SkillRevise: Improving LLM-Authored Agent Skills via Trace-Conditioned Skill Revision
- ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents
- Understanding Code Agent Behaviour: An Empirical Study of Success and Failure Trajectories
- Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills
- A Self-Evolving Framework for Efficient Terminal Agents via Observational Context Compression
- Skills-Coach: A Self-Evolving Skill Optimizer via Training-Free GRPO
- SkillGrad: Optimizing Agent Skills Like Gradient Descent
- Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward
- SkillOpt: Executive Strategy for Self-Evolving Agent Skills
- SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories
- SkillEvolver: Skill Learning as a Meta-Skill
- CoEvoSkills: Self-Evolving Agent Skills via Co-Evolutionary Verification
Related papers
- Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
- GitSkills: A Dataset of Agent Skills on GitHub
- SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
- PackMonitor: Enabling Zero Package Hallucinations Through Decoding-Time Monitoring
- IntentCoding: Amplifying User Intent in Code Generation
- Incentives and Outcomes in Bug Bounties