《Cursor文檔》-Glm 5 3 Flash
GLM 5.3 Flash is the smaller, faster member of Z.ai's GLM 5.3 family. It supports image inputs, a 1M token context window, and three reasoning levels (Low, High, and Max), with High as the default. Add it to the model picker from Cursor Settings Models . For higher quality on harder tasks, use GLM 5.3.
閱讀全文《Cursor文檔》-Glm 5 3
GLM 5.3 is Z.ai's latest open weight coding model and the successor to GLM 5.2. It supports a 1M token context window and three reasoning levels (Low, High, and Max), with High as the default. Add it to the model picker from Cursor Settings Models . For a faster, lower cost option from the same family, see GLM 5.3 Flash.
閱讀全文《Cursor文檔》-Gemini 4 Argon
Gemini 4 Argon is Google's frontier Gemini model. Use it for complex, multi step coding work that benefits from deeper reasoning and a large context window.
閱讀全文《Cursor文檔》-Gemini 3 5 Flash Lite
This model is used only by the Auto review classifier inside Cursor. It isn't selectable as a chat or agent model.
閱讀全文《Cursor文檔》-Claude Sonnet 5 5
Claude Sonnet 5.5 is Anthropic's latest medium tier model and the successor to Sonnet 5. It scores 55.5% on CursorBench at max effort, second only to Opus 5.5. It keeps Sonnet 5's input and output rates and halves the cache read rate. Add it to the model picker from Cursor Settings Models .
閱讀全文《Cursor文檔》-Claude Opus 5 5
Claude Opus 5.5 is Anthropic's latest Opus model and replaces Opus 5. It sets a new high score on CursorBench, ahead of Fable 5.1, and costs 20% less than Opus 5 per token. We recommend the high thinking variant for the best results.
閱讀全文《Cursor文檔》-Claude Haiku 5 5
Claude Haiku 5.5 is the fastest model in Anthropic's Claude 5.5 family and the successor to Claude 4.5 Haiku. It scores 48.4% on CursorBench at max effort, on the same cost performance curve as Sonnet 5.5 and Opus 5.5, at a tenth of Haiku 4.5's per token price for requests up to 100k input tokens. Add it to the model picker from Cursor Settings Models .
閱讀全文《Cursor文檔》-Deploy Grok Bot to your organization
Your IT team installs the Grok Bot desktop app on member devices with the tools it already uses for other software: download a versioned installer, push it, and decide when devices move to a new release. Dashboard controls for turning Grok Bot on and scoping access are on Grok Bot for Teams and Enterprise, and deploying the Cursor editor is covered in Deployment patterns.
閱讀全文《Cursor文檔》-Manage Grok Bot computers
Each member of your team gets one hosted computer where every Bot they run does its work. Manage Member Computers and Data , under Security & Compliance on the Grok Bot page of the Cursor dashboard, lets organization admins recreate or terminate computers across many members at once, or delete all of a member's Grok Bot data. Terminate Inactive Computers , under Bot Computers , turns on automatic termination for computers that go 30 days without use. For the rest of the admin controls, see Grok Bot for Teams and Enterprise.
閱讀全文《Cursor文檔》-Tenancy Restriction
Your network can limit which Cursor organizations and teams a request may use. Do that by adding one HTTP header on traffic to Cursor. Cursor never sets this header, and the dashboard has no switch for it. If the header is absent, Cursor applies no restriction from this mechanism.
閱讀全文《Cursor文檔》-Origin Migration API
Origin is in Early Beta and subject to change. Review the OpenAPI specification when updating an integration.
閱讀全文《Cursor文檔》-Origin MCP tool reference
Every tool the Origin MCP server at https://api.origin.cursor.com/mcp lists, in listing order. Parameters are the JSON Schema clients receive from tools/list.
閱讀全文《Cursor文檔》-Origin API Changelog
The Origin API is in Early Beta and subject to change. Review the OpenAPI specification when updating an integration.
閱讀全文陳庚-內容策略：把保險顧問定位轉成可執行的內容打法
陳庚-內容策略（@user_88e00a20/chengeng-content-strategy，v1.0.2）是 SkillHub 社區內容創作類 Skill，面向已有定位與人設的保險顧問。它不重做人設，而是判斷內容目標、區分增長/篩選/信任/承接/服務顯化，輸出內容主線、渠道分工、選題池、7天與30天發佈節奏、低壓力承接及風險表達檢查；內置輸入校驗、重試與可選質量門。本文介紹能力邊界、skillhub 安裝命令（含 Cursor skills 目錄）及典型觸發方式，並附官方頁面與 API 地址。
閱讀全文《Cursor文檔》-Cursor Origin API
Early Beta reference for the Cursor Origin API. Origin is Cursor's code forge; its public REST API lets apps and tools work with Origin repositories, commits, checks, pull requests, apps, and installations.
閱讀全文《Cursor文檔》-Projects
A Project takes on a larger body of work, such as a feature, a migration, or a full app. You direct it by chatting with its coordinator agent. The coordinator doesn't write code itself. It plans the work, delegates it to agents that write the code, and brings the finished work back to you to check. Projects live in the left hand nav of the Agents Window and run on Cloud Agents.
閱讀全文《Cursor文檔》-Legacy Cursor plans
Legacy Cursor plans are the plans Cursor sold before The X Subscription: Hobby, Start, Pro, Pro+, Ultra, and Teams with Standard and Premium seats. They keep working as they are. New customers start on The X Subscription, and this page is for anyone still on a legacy plan who wants to know how it works and how it compares.
閱讀全文一個會自進化的日誌分析工具：確定性規則與 AI 推理結合的 Agent Skill
介紹 SkillHub 社區 Skill @user_00c9b356/buddy-log-analyzer（v1.0.0，it-ops-security）。官方定位爲日誌智能分析工具，採用 mini-swe-agent 極簡理念：確定性規則與 AI 推理結合。文章說明來源、requires_api_key 標籤、SkillHub CLI 安裝命令及各 Agent skills 目錄要點，並說明公開頁未附完整 SKILL.md 時如何按通用 Skill 模式在會話中使用。
閱讀全文dsh-thinking-guard：純思考空轉熔斷器
`dsh-thinking-guard` 插件通過三重熔斷機制（超時、容量、重複）防止 Agent 進入“只思考、零產出”的無限循環。該插件監聽 `agent/assistant-stream` 事件，在純思考、內容超限或模式退化時自動觸發 `agent.cancel` 中斷任務。支持自定義參數配置及環境變量覆蓋，已修復相關代碼缺陷。
閱讀全文PPT分析與製作規劃助手：把零散素材整理成可執行的 PPT 大綱
PPT分析與製作規劃助手（@user_44beaa18/ppt-analysis-designer，v1.0.0）是 SkillHub 社區 office-efficiency 類 Skill，無需單獨 API Key。它分析文本、表格、數據、圖片或主題，輸出結構化 PPT 內容大綱，並覆蓋生成大綱、優化結構、答辯講稿與單頁版式建議等常見表述。文章說明安裝 CLI、用 skillhub install 裝入 Cursor 等 skills 目錄，並給出可復現的對話示例與適用邊界。
閱讀全文課題結題報告寫作模板：把教科研報告寫成「研究結論」而非活動流水賬
SkillHub 社區 Skill「課題結題報告寫作模板」（@user_777749ea/001-new，v1.0.1）面向教育科研課題的研究報告與結題報告寫作。在用戶提及結題、研究報告、教科研報告等場景時觸發，提供標準化學術結構、分步寫作指導與質檢清單，幫助從「做了什麼」上升到「研究出了什麼」。本文介紹其定位、核心能力、基於 skillhub CLI 的安裝方式（含 Cursor/Codex 等 skills 目錄），以及可復現的對話用法與使用注意。
閱讀全文WorkBuddy Auto Checkin：用 wbipc 複用客戶端會話做桌面端自動簽到
WorkBuddy 自 2026-09-27 起本地加密憑據，明文 token 調 HTTP API 的簽到方式已失效。SkillHub 社區 Skill「WorkBuddy Auto Checkin」（v4.0.1，@user_24b4fab8/workbuddy-auto-checkin-macos）指導 Agent 用本地 wbipc RPC 橋複用客戶端登錄會話，完成 Buddy 加油站每日能量包領取，並覆蓋腳本重建、launchd 排障與發佈 Skill 等場景。安裝需 skillhub CLI，將技能裝入對應 Agent 的 skills 目錄；無需 API Key，主要面向 macOS 桌面 WorkBuddy 用戶。
閱讀全文