做 LLM Agent 开发的团队终于有了一个系统化的技能进化方法，不用再靠试错和启发式反思——Bayesian-Agent 用后验概率指导优化，效果显著且可审计，建议直接看论文和代码。
Bayesian-Agent 是一个跨框架的 LLM Agent 技能优化框架，它将可复用的技能和 SOP 视为关于冻结模型在特定提示、上下文和环境下能否成功的假设。该框架记录验证过的轨迹证据，维护每个技能的特征条件分类后验，并根据后验状态执行修补、拆分、压缩、退役和探索等操作。在 deepseek-v4-flash 上，增量修复将 SOP-Bench 从 80% 提升至 95%，Lifelong AgentBench 从 90% 提升至 100%，RealFin-Bench 从 45% 提升至 65%。该框架还支持 GenericAgent、mini-swe-agent 和 Claude Code 等后端，表明 Agent 技能进化应视为后验引导的优化，而非未校准的提示积累。代码已开源。
Bayesian-Agent: Posterior-Guided Skill Evolution for LLM Agent Harnesses
LLM agents increasingly rely on external inference conditions: prompts, tools, memory, SOPs, skills, and harness feedback. These assets can improve task execution without changing model weights, but they are often revised by heuristic reflection or by reusing observed successes and failures as if counts alone were reliable belief. We introduce \textbf{Bayesian-Agent}, a native and cross-harness framework that treats reusable skills and SOPs as hypotheses about whether a frozen model will succeed under a particular prompt, context, and harness environment. Bayesian-Agent records verified trajectory evidence, maintains a feature-conditioned categorical posterior over each skill, and maps posterior state into inspectable actions such as patch, split, compress, retire, and explore. Model-facing prompts receive executable guardrails and failure-mode patches, while posterior summaries remain available for audit. With \texttt{deepseek-v4-flash}, incremental repair improves SOP-Bench from 80\% to 95\%, Lifelong AgentBench from 90\% to 100\%, and RealFin-Bench from 45\% to 65\%. We further evaluate Bayesian-Agent's native backend and optional GenericAgent, mini-swe-agent, and Claude Code backends. The results include positive, negative, saturated, and case-study settings, suggesting that agent skill evolution is best viewed as posterior-guided harness optimization rather than uncalibrated prompt accumulation. The source code is available at https://github.com/DataArcTech/Bayesian-Agent.