让AI代理像操作系统一样运行，用Markdown文件统一管理应用界面与操作。
String: An Agentic OS Where Every App Is a Markdown File
- 将应用定义为Markdown文档，通过/open和/act指令实现智能体交互。
- 在87项任务中成功率接近主流模型，但减少33.5%的调用Token消耗。
- 支持网页、应用、终端等多端统一接口，适合构建轻量级智能体应用。
大语言模型代理已成为新型软件使用者，但其操作界面仍为人类设计：页面适合人浏览（可跳读忽略），工具接口针对程序编写，却无需承担未调用定义的开销。而代理无法享受这些便利：它每轮必须重读全部内容，并为每次调用支付代价。我们提出String——一个开源运行时，将此问题视为操作系统层面的设计挑战。工具知识从代理上下文中移出，进入公共层，按需以Markdown形式逐视图返回。单个SFMD（String-Flavored Markdown）文档即可声明应用的视图、类型化动作、导航路径与凭据。运行时负责发现、验证、执行、状态维护及密钥管理，仅需两个核心命令：/open查看与/act执行。网页与应用实为同一架构的不同渲染：一个SFMD站点既向浏览器提供样式化的HTML，也向代理提供原始文档，使同一语法覆盖应用、文件、终端与旧版HTML，无需每站集成。视图设计为部分暴露，且流程具有因果性：提前披露一层信息最多导致23分准确率下降，正确分层则使错误动作选择从28%降至2%。权限随来源而定：远程页面可调用HTTP但不可访问终端，用户输入绝不扩展存储的秘密。在一项包含87个任务的基准测试中，每个任务搭配定制技能，将这些流程转化为按需部署的String应用后，六种不同规模的模型表现相当（领先+1.3个百分点），完成回合中调用令牌减少33.5%，且界面始终维持在53个令牌不变，无论目录规模如何。本文报告了设计、评估及三个月生产使用经验。
原文摘要 · Abstract (English)
LLM agents have become a new class of software user, but every surface they work through was designed for someone else. Pages are built for human eyes, which can skim and ignore; tool schemas for programs, which pay nothing to carry definitions they never call. An agent has neither luxury: it re-reads, and pays again for, everything it is shown on every turn. We present String, an open-source runtime that gives this user an interface of its own and treats the job as an operating-systems problem. Tool knowledge moves out of the agent's context and into a common layer that renders it back one view at a time as Markdown. A single SFMD (String-Flavored Markdown) document declares an application's views, typed actions, navigation, and credentials, and the runtime handles discovery, validation, execution, state, and secrets behind two core verbs: /open to see and /act to do. Web and app turn out to be two renderings of one architecture: an SFMD site serves styled HTML to browsers and the raw document to agents, so one grammar reaches apps, files, shells, and the web, even legacy HTML, with no per-site integration. Views stay partial by design, and the staging is causal: disclosing one tier of detail a single turn too early costs up to 23 accuracy points, while proper staging drops wrong-action selection from 28% to 2%. Privilege follows provenance: a remote page may call HTTP but never the shell, and caller-supplied text never expands a stored secret. On an 87-task benchmark that pairs each task with curated skills, operationalizing those procedures as on-demand String apps yields comparable aggregate success across six models from frontier to small (+1.3pp) while using 33.5% fewer tokens among completed episodes, and the resident interface stays a constant 53 tokens at any catalog size. We report the design, the evaluation, and what three months of production use taught us.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。