元访问的威力：以 LLM 聊天界面为例探讨未来软件中的 Meta Access
The Power of Meta Access in the Software of the Future --- Demonstrated on LLM Chat UIs and Chat AI
AI 导读
arXiv 论文提出“元访问”概念：让 LLM 助手获得对界面本身的结构化、带权限的视图，并给出从功能教学、元数据访问到检索、确认修改与完全编排的能力阶梯。作者在 Chat AI 客户端的 Meta Chat 原型中落地该方案，支持从摘要、设置和相关消息构建有界上下文、确认可逆更改与教程库，同时指出 grounding、隐私和提示词注入等风险。
正文 · 原文
Abstract:Chat remains the default surface for many large language model (LLM) tasks, even as agentic tools such as OpenCode excel at coding and automation. Yet the model sees the current dialogue but not the workspace around it. We explore meta access---giving the assistant a structured, permissioned view of the interface itself and propose a ladder of levels: feature tutoring; metadata, settings, and preference access; retrieval over summaries and messages, confirmed mutation, and full orchestration. Meta access can turn a chat client from many screens into one entry point for finding, navigating, organising, and operating it, while raising risks around grounding, privacy, prompt injection, and trust. General-purpose agentic harnesses are runtimes; because they can change their own settings, meta access becomes reflexive and needs a separate control-plane boundary. We ground this in the Meta Chat prototype of the Chat AI client. It builds bounded context from, e.g., summaries, settings, and relevant messages; links references to conversations, topics, and personas; supports confirmed, reversible changes; offers a browsable tutorial library; and recommends personas as user-started conversations. By exposing application state in structured form, meta access can extend screen readers from widget descriptions to semantic workspace operations. Meta access thus turns chat from a window onto one conversation into a control layer for the whole workspace, extending chat toward agentic capability without surrendering user control. As assistants become embedded in IDEs, dashboards, office suites, and operating systems, AI-era interfaces should expose a structured, permissioned model of their state and capabilities, ideally by opening their existing metamodels through a debugger-like interface.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.08428 [cs.AI] |
| (or arXiv:2611.08428v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.08428 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Julian Kunkel [view email]
[v1]
Thu, 5 Nov 2026 18:22:24 UTC (1,787 KB)
来源：arXiv cs.AI · arxiv.org