揭秘代理技能：为何有效直至失效
原标题：Paper page - Demystifying Agent Skills: Why They Work-Until They Don't
AI 摘要
该论文通过受控实验和轨迹分析，研究了技能（Skills）增强LLM代理的机制。研究发现，技能主要通过程序锚定（procedural anchoring）稳定执行，而非注入缺失知识，程序锚定占技能案例的65.7%，而显式知识注入仅占4.5%。检索是独立瓶颈，当技能池从5增至100时，实际使用精度从29.6%降至3.3%。技能在脆弱假设、不兼容上下文或适应不足时失效。研究提出了技能使用模式的分类法，并推动评估超越聚合成功率。
正文节选
Demystifying Agent Skills: Why They Work-Until They Don't Abstract Skills enhance LLM agents primarily by stabilizing execution through procedural anchoring rather than injecting missing knowledge, though retrieval bottlenecks and brittle assumptions limit their effectiveness. Skills have emerged as a practical and effective approach for enhancing LLM agents at inference time through structured packages of knowledge. However, existing evaluations largely measure whether skills improve aggregated