Cover Image

More Skills, Worse Agents? Skill Shadowing Degrades Performance When Expanding Skill Libraries

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Song, Hongwen, Song, Wei
Format:	Preprint
Published:	2026
Subjects:	Software Engineering Artificial Intelligence Applications
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

SWE-Skills-Bench: Do Agent Skills Actually Help in Real-World Software Engineering?
by: Han, Tingxu, et al.
Published: (2026)

Skill Drift Is Contract Violation: Proactive Maintenance for LLM Agent Skill Libraries
by: Fan, Linfeng, et al.
Published: (2026)

When Safe Skills Collide: Measuring Compositional Risk in Agent Skill Ecosystems
by: Wang, Su, et al.
Published: (2026)

SkillSmith: Compiling Agent Skills into Boundary-Guided Runtime Interfaces
by: Xu, Duling, et al.
Published: (2026)

SkillMOO: Multi-Objective Optimization of Agent Skills for Software Engineering
by: Gong, Jingzhi, et al.
Published: (2026)

ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents
by: Lu, Zijian, et al.
Published: (2026)

Skilled AI Agents for Embedded and IoT Systems Development
by: Li, Yiming, et al.
Published: (2026)

Scaling Coding Agents via Atomic Skills
by: Ma, Yingwei, et al.
Published: (2026)

Contractual Skills: A GovernSpec Design Framework for Enterprise AI Agents
by: Liu, Ting
Published: (2026)

Skill Discovery for Software Scripting Automation via Offline Simulations with LLMs
by: Xu, Paiheng, et al.
Published: (2025)

Kimi-Dev: Agentless Training as Skill Prior for SWE-Agents
by: Yang, Zonghan, et al.
Published: (2025)

Library Drift: Diagnosing and Fixing a Silent Failure Mode in Self-Evolving LLM Skill Libraries
by: Zhang, Xing, et al.
Published: (2026)

HarnessAPI: A Skill-First Framework for Unified Streaming APIs and MCP Tools
by: Jose, Edwin
Published: (2026)

TransCoder: Towards Unified Transferable Code Representation Learning Inspired by Human Skills
by: Sun, Qiushi, et al.
Published: (2023)

Capability-Driven Skill Generation with LLMs: A RAG-Based Approach for Reusing Existing Libraries and Interfaces
by: da Silva, Luis Miguel Vieira, et al.
Published: (2025)

Agent Skills in the Wild: An Empirical Study of Security Vulnerabilities at Scale
by: Liu, Yi, et al.
Published: (2026)

SkillOps: Managing LLM Agent Skill Libraries as Self-Maintaining Software Ecosystems
by: Pu, Hongji, et al.
Published: (2026)

BugPilot: Complex Bug Generation for Efficient Learning of SWE Skills
by: Sonwane, Atharv, et al.
Published: (2025)

SkillForge: Forging Domain-Specific, Self-Evolving Agent Skills in Cloud Technical Support
by: Liu, Xingyan, et al.
Published: (2026)

SkillReducer: Optimizing LLM Agent Skills for Token Efficiency
by: Gao, Yudong, et al.
Published: (2026)

Bridging the Skills Gap: A Course Model for Modern Generative AI Education
by: Bardach, Anya, et al.
Published: (2025)

One Model, Many Skills: Parameter-Efficient Fine-Tuning for Multitask Code Analysis
by: Akli, Amal, et al.
Published: (2026)

Enhancing Debugging Skills with AI-Powered Assistance: A Real-Time Tool for Debugging Support
by: Artser, Elizaveta, et al.
Published: (2026)

Skills as Verifiable Artifacts: A Trust Schema and a Biconditional Correctness Criterion for Human-in-the-Loop Agent Runtimes
by: Metere, Alfredo
Published: (2026)

Beyond Formal Semantics for Capabilities and Skills: Model Context Protocol in Manufacturing
by: da Silva, Luis Miguel Vieira, et al.
Published: (2025)

SkillClone: Multi-Modal Clone Detection and Clone Propagation Analysis in the Agent Skill Ecosystem
by: Zhu, Jiaying, et al.
Published: (2026)

SkillCraft: Can LLM Agents Learn to Use Tools Skillfully?
by: Chen, Shiqi, et al.
Published: (2026)

EffiSkill: Agent Skill Based Automated Code Efficiency Optimization
by: Wang, Zimu, et al.
Published: (2026)

When Agents go Astray: Course-Correcting SWE Agents with PRMs
by: Gandhi, Shubham, et al.
Published: (2025)

ContraFix: Agentic Vulnerability Repair via Differential Runtime Evidence and Skill Reuse
by: Liu, Simiao, et al.
Published: (2026)

SkillProbe: Security Auditing for Emerging Agent Skill Marketplaces via Multi-Agent Collaboration
by: Guo, Zihan, et al.
Published: (2026)

When Neural Code Completion Models Size up the Situation: Attaining Cheaper and Faster Completion through Dynamic Model Inference
by: Sun, Zhensu, et al.
Published: (2024)

RUM: Rule+LLM-Based Comprehensive Assessment on Testing Skills
by: Wang, Yue, et al.
Published: (2025)

When the Specification Emerges: Benchmarking Faithfulness Loss in Long-Horizon Coding Agents
by: Yan, Lu, et al.
Published: (2026)

Debug2Fix: Can Interactive Debugging Help Coding Agents Fix More Bugs?
by: Garg, Spandan, et al.
Published: (2026)

When Context Hurts: The Crossover Effect of Knowledge Transfer on Multi-Agent Design Exploration
by: Vigraham, Saranyan
Published: (2026)

Exploring Autonomous Agents: A Closer Look at Why They Fail When Completing Tasks
by: Lu, Ruofan, et al.
Published: (2025)

When AI Teammates Meet Code Review: Collaboration Signals Shaping the Integration of Agent-Authored Pull Requests
by: Nachuma, Costain, et al.
Published: (2026)

RedCode: Risky Code Execution and Generation Benchmark for Code Agents
by: Guo, Chengquan, et al.
Published: (2024)

ProcCtrlBench: Evaluating Process-Level Defects and Control Preservation in LLM Coding Agents
by: He, Jiawei, et al.
Published: (2026)