Zep 推出 Agent Skills 服务：从验证轨迹自动编译智能体技能
Learning by Doing: Autonomously Compiling and Curating Agent Skills from Verified Traces
Zep 把智能体的成功操作记录编成可复用的技能包，弱模型直接拿来用通过率最多涨 18.8 点，还不用训练，做 Agent 的可以看看这套流程。
Zep Agent Skills 是一项面向生产环境智能体的程序性记忆服务，把每次任务执行记录为带事件序列、结果和验证强度的 Trace。一个由 Reflector 和 Curator 组成的编译器只用通过验证的 Trace 生成版本化的 Skills，并通过扫描和准入门槛控制每个 Skill 版本的启用。该格式源自 Anthropic 发起并捐给 Agentic AI Foundation 的开放标准，宿主可通过 API 或 MCP 加载。在 τ³-bank banking、AppWorld 和 EnterpriseOps-Gym 上，教师模型 Trace 编译的 Skills 让三个较弱的执行模型在不同推理档位下通过率提升 4.5 到 18.8 点。在 ALFWorld 上，小模型运行编译的 Skills 收益达到甚至超过先前工作里持续微调的收益，且无需训练。
Learning by Doing: Autonomously Compiling and Curating Agent Skills from Verified Traces
Memory systems for production LLM agents mostly store semantic memory (facts about users and domains) or episodic memory (earlier interactions). Research systems such as Trace2Skill, WikiSkill, and SkillRevise learn procedures from agent runs, package them as Skills, and evaluate them on benchmarks. We describe Zep Agent Skills, a production service that gives deployed agents this procedural memory. The service uses only verified attempts as evidence, cites the supporting events for each field of a Skill, and governs the admission and distribution of each Skill version. The service records each task attempt as a Trace with ordered events, an outcome, and a verification strength. A compiler with separate Reflector and Curator roles turns verified Traces into versioned Skills. Scans and admission gates control which candidates become active, and verified use outcomes change the search rank of each Skill. The Agent Skills format is an open standard that Anthropic originated and donated to the Agentic AI Foundation, so any host of the format can load a compiled Skill. Agents retrieve Skills through lexical search with progressive disclosure, over an API or MCP. We evaluate Zep Agent Skills on $τ^3$-bench banking, AppWorld, and EnterpriseOps-Gym. Skills compiled from the Traces of a teacher model increase the pass rate of three weaker executor models on the same tasks, at every reasoning effort, by 4.5 to 18.8 points. Most of the gain is on the tasks that the teacher solved. On ALFWorld, Skills compiled from the runs of a small model give a gain similar to or larger than the gain that prior work reports for continuous fine-tuning, with no training.
