4.0.2

Release Notes: v4.0.2 - Codebase Audit, Dead Code Pruning & SmartComposer Decoupling / 4.0.2 工程代码深度审查、死代码清理与上游彻底脱耦

🌐 English

• Dead Code & Bloat Dependencies Pruned:
  • Uninstalled heavyweight unused dependencies (react-syntax-highlighter, vscode-diff), reducing the final bundle size from 0.44MB to 0.41MB.
  • Deleted obsolete chat view components (MarkdownCodeComponent, SyntaxHighlighterWrapper) and switched to Obsidian's native MarkdownRenderer for standard code blocks.
  • Eliminated legacy <smtcmp_block> parsing logic, focusing strictly on <think> blocks for better regex efficiency.
• Main Class (God Class) Refactoring:
  • Extracted the massive settings validation, legacy provider filtering, and dynamic model enforcement logic from main.ts into a dedicated SettingsManager for cleaner architecture (SOLID).
• Version Bump: Updated to version 4.0.2.

🇨🇳 中文说明

• 剔除庞大废弃依赖与无效代码（极致瘦身）：
  • 彻底卸载了早已过时且庞大的依赖（react-syntax-highlighter、vscode-diff），将打包产物体积由 0.44MB 进一步压榨至 0.41MB。
  • 清理了与上述依赖绑定的旧版聊天渲染组件，全面拥抱 Obsidian 原生的代码块渲染器（支持原生复制按钮）。
  • 去除原先兼容上游的 <smtcmp_block> 正则解析，精简解析流并专攻 <think> 思考标签，降低运行时开销。
• 化解入口类 God Class 困境（架构升级）：
  • 将原先堆积在 main.ts 中的数百行硬编码配置清洗、虚假服务商过滤与容错逻辑全盘抽离至独立的 SettingsManager，使工程架构更符合高内聚低耦合标准。
• 版本发布：版本号升级为 4.0.2。

---

1.0.0

Release Notes: v1.0.0 - Official Milestone Release / 1.0.0 正式里程碑版本

🌐 English

• Vault Chat & RAG Quick Setup Guided Workflow:
  • Interactive setup modal automatically guides users when initiating vault-wide knowledge base chat without active RAG or embedding model configuration, seamlessly continuing chat upon saving.
  • Seamless localization resolution: modal strictly honors system language (auto / Obsidian locale) across titles, notice toasts, and settings controls.
• Visual Design & Typography Refinement:
  • Dual Action Button Consistency: "Article Chat" and "Vault Chat" buttons now share identical refined accent borders and subtle backgrounds matching Obsidian theme palettes.
  • Balanced User Bubble Styling: Removed saturated user bubble background, harmonizing typography, border, and line height with assistant responses.
• RAG & Knowledge Base Enhancements:
  • Exclusion of Vault System Folders: Enforced default exclusion of non-note directories (.obsidian, .aider, .aide, .trash, .git, .smart-env) to safeguard indexing integrity.

🇨🇳 中文说明

• 全知识库对话引导弹窗与 RAG 交互闭环：
  • 前台发起全知识库检索对话时，若 RAG 未启用或未选择嵌入模型，弹出就地配置引导窗口，保存后直接自动无缝提交对话，无需重复点击。
  • 多语言解析跟随系统：彻底修复在 auto（跟随系统）语言下弹窗偶现英文的问题，界面、提示词与按钮纯正本地化。
• 界面视觉风格与排版深度优化：
  • 对话操作按钮模式统一：“文章对话”与“全知识库对话”统一采用精致强调边框（var(--interactive-accent)）与微底色，视觉认知清晰协调。
  • 用户消息气泡去除突兀底色：用户消息转为无底色细边框卡片，版式、字重与行高与 AI 回复完全统一，沉浸感大幅增强。
• RAG 知识库与文本切分精细化：
  • 系统目录底层默认强制排除：知识库检索黑名单默认强制排除 .obsidian、.aider、.trash、.git 等典型非库内容系统目录，保障扫描精准高效。

---

0.13.3

Release Notes: v0.13.3 - Obsidian Community Review Compliance & Security Hardening / 官方社区审核合规与安全加固

🌐 English

• Obsidian Official Community Review Compliance:
  • Eliminated root-cause linter errors (react-hooks/exhaustive-deps) by annotating explicit justifications in accordance with Obsidian review guidelines.
  • Standardized window/timer API calls across the plugin for seamless compatibility with Obsidian Popout multi-window contexts.
  • Hardened asynchronous lifecycle and UI event handlers (ObsidianButton, ObsidianToggle, ObsidianDropdown, ObsidianTextInput, ObsidianTextArea, etc.) with safe void Promise resolution, completely resolving floating promise warnings.
• TypeScript Strong Typing & Quality Refactoring:
  • Refactored MentionableBadge, mcpManager, gemini, and migration scripts to eliminate unsafe any casts and member accesses.
  • Enhanced openMarkdownFile signature to support both TFile and file path strings seamlessly.
• Dependency Security Patching:
  • Upgraded vulnerable dependencies including uuid (^11.1.1) and completed full npm audit fix.
  • Cleaned up obsolete properties and debug logs across adapters.
• Version Bump: Updated to version 0.13.3.

🇨🇳 中文说明

• Obsidian 官方社区审核全项合规化：
  • 修复硬阻断 ESLint 指令注释，补充明确规则声明与业务场景阐述，达到官方审查最高标准。
  • 标准化定时器与动画帧 API，显式绑定至独立 window 上下文，彻底保障 Popout 多窗口环境下的运行稳定性。
  • 深度重构底层基础交互组件（按钮、开关、下拉框、文本框）及业务组件生命周期的异步回调，统一采用安全的浮动 Promise 消费机制，消除所有 Promise 悬空警告。
• TypeScript 强类型重构与代码质量提升：
  • 全面消除 MentionableBadge、MCP 服务管理、Gemini 适配器及历史配置迁移脚本中的 any 与不安全类型断言，增强类型健壮性。
  • 增强 openMarkdownFile 工具方法，无缝支持 TFile 对象与字符串路径双重传参。
• 依赖安全加固与代码瘦身：
  • 升级 uuid 等核心依赖至安全版本 ^11.1.1，完成完整依赖安全审计。
  • 清理适配器中废弃的 system_fingerprint 赋值与冗余日志。
• 版本更新：版本号升级为 0.13.3。

---

0.13.2

Release Notes: v0.13.2 - Template Modals Blank Bugfix & Stability / 模板弹窗空白修复与稳定性优化

🌐 English

• Fix Template Modals Blank/Crash: Resolved an issue where opening the template selection modal from the chat input or clicking "Add/Edit Prompt Template" in Settings would render a completely blank modal due to missing SettingsProvider context in React modals.
• Resilient Multi-Language Fallback: Hardened useI18n() hook with a safe fallback mechanism to prevent runtime crashes when rendered outside the global settings context.
• Robust Template Loading: Implemented interval polling retry logic when loading existing template nodes into the Lexical editor during template editing.
• Version Bump: Updated to 0.13.2.

🇨🇳 中文说明

• 彻底修复模板相关弹窗内容空白崩溃问题：修复了在前台聊天输入框点击“提示词模板”图标、以及在后台设置中点击“添加提示模板”或“编辑”已有模板时，因 React 模态框缺失 SettingsProvider 导致 useI18n 抛出未捕获异常并引起白屏的缺陷。
• 国际化多语言容错与安全降级：强化 useI18n() 钩子，增加安全 Context 查询与默认语种降级，杜绝任何独立组件或弹窗因语种上下文未就绪而导致崩溃。
• 模板编辑加载时序增强：在编辑已保存模板时，增加 Lexical 编辑器节点挂载状态的轮询重试机制，确保复杂模板内容 100% 稳定呈现。
• 版本更新：版本号升级为 0.13.2。

---

0.11.0

Release Notes: v0.11.0 - Obsidian Aider Official Release / 正式发布

🌐 English

• Interactive In-Place Question Editing: Double-clicking on any user message bubble directly activates in-place editing mode, allowing fast query modification and seamless re-generation.
• Unified Directory Migration: Standardized the vault configuration and storage directory from .aide to .aider, with automatic background migration on startup.
• Streamlined MCP Tool Interface: Removed internal sandbox security banners from the MCP settings tab to deliver a cleaner, distraction-free configuration experience.
• Robust Long Filename Handling: Resolved potential Windows path length overflow (>255 chars) during chat session and template title encoding, ensuring reliable multi-language vault synchronization.
• Version Bump: Updated to version 0.11.0.

🇨🇳 中文说明

• 双击编辑历史提问：前端对话界面中，双击任意用户消息气泡即可直接进入就地编辑状态，支持快速修改与重新生成。
• 项目存储目录规范化：存储与配置目录全面规范为 .aider，启动时自动平滑迁移历史 .aide 目录，无缝兼容老版本数据。
• MCP 界面极简精简：移除 MCP 工具面板中冗余的“纯净沙箱安全模式”提示，为普通用户提供更清爽直观的交互界面。
• 中文与长文件名防溢出优化：重构对话历史与模板命名逻辑，彻底解决 Windows 系统下因 URL 编码导致的长文件名路径超限（>255 字符）报错问题。
• 版本发布：版本号升级为 0.11.0。

---

0.10.0

Release Notes: v0.10.0 - Obsidian Aider Official Release / 正式发布

🌐 English

• Full Cross-Platform & Mobile Readiness: Completely decoupled from Node.js subprocesses and direct filesystem access (child_process and fs reduced to 0). Runs purely within Obsidian's official sandbox, fully compliant with the latest Obsidian Community Plugin security review standards, and seamlessly supported across both Desktop and iOS/Android mobile devices.
• Extreme Lightweight Optimization: Bundle size further decreased from ~2.1 MB to 0.43 MB (over 70% reduction), ensuring instantaneous startup and zero lag on your vault.
• Native Zero-Config Built-in Tools: 5 native tools (Bing Web Search, Webpage Content Fetch, Global Weather Forecast, arXiv Academic Search, Accurate System Time) running purely on Obsidian's cross-platform network API without extra configuration or API keys.
• Comprehensive Multi-Provider Ecosystem: 4 core providers (OpenAI, DeepSeek, OpenRouter, SiliconFlow) enabled by default for out-of-the-box ease, with full support for adding Anthropic (Claude), Google Gemini, xAI (Grok), Mistral AI, Perplexity, Groq, ModelScope, Azure OpenAI, local offline models (Ollama, LM Studio), and any OpenAI-compatible custom endpoints.
• Version Bump: Updated to version 0.10.0.

🇨🇳 中文说明

• 全平台合规与移动端无缝支持：彻底移除 Node.js 子进程与直接文件系统访问依赖（child_process 与 fs 引用归零），完全基于 Obsidian 官方沙箱环境与跨平台网络接口运行，完美通过官方插件中心最新安全审核，全方位支持桌面端与 iOS / Android 移动设备。
• 极致轻量化精简：打包体积由 ~2.1 MB 大幅缩减至 0.43 MB（瘦身超 70%），秒开加载，不占用多余内存，不拖慢笔记库检索。
• 原生内置工具增强：5 大开箱即用原生工具（国内必应搜索、网页抓取转 Markdown、全球实时天气预报、arXiv 学术文献检索、当前系统精准时间）原生极速响应，免 API Key、免额外环境配置。
• 全方位多模型服务商矩阵：默认预置启用 4 家核心服务商（OpenAI、DeepSeek、OpenRouter、SiliconFlow），并全面支持一键添加配置 Anthropic (Claude)、Google Gemini、xAI (Grok)、Mistral、Perplexity、Groq、魔搭社区 (ModelScope)、Azure OpenAI、本地离线大模型 (Ollama、LM Studio) 以及任意兼容 OpenAI 规范的自建与中转服务。
• 版本发布：版本号升级为 0.10.0。

---

0.9.0

Release Notes: v0.9.0 - Obsidian Aider Official Release / 正式发布

🌐 English

• Default Language: English is now the default interface and system prompt language.
• Global System Prompt: Standardized to English, brand name updated to "Aider".
• Runtime Profile: Defaulted to Free Tier / Eco mode (eco) with rate-control discipline and token efficiency.
• Retained Providers: Core lineup streamlined to 4 primary providers: OpenAI, DeepSeek, OpenRouter, and SiliconFlow.
• Default Chat Model: Set to 1 model by default: SiliconFlow's deepseek-ai/DeepSeek-V4-Flash.
• Version Bump: Bumped to version 0.9.0.

🇨🇳 中文说明

• 默认语言：项目语言默认为英文（支持随时在设置中切回中文）。
• 全局系统提示词：改为英文版，并将提示词中的 Aide 规范为 Aider。
• 服务层级：默认为免费层模式（🌱 Eco Mode，严格控流、精简 Token、防 429）。
• 模型服务商：精简保留 4 家核心模型服务商（OpenAI、DeepSeek、OpenRouter、SiliconFlow）。
• 对话模型：默认配置 1 个对话模型（SiliconFlow 的 deepseek-ai/DeepSeek-V4-Flash）。
• 版本发布：版本号升级为 0.9.0。

---

v0.8.3

Release Notes: v0.8.3 - Obsidian Aide Official Release / 正式发布

🌐 English

🎉 Welcome to Obsidian Aide!

Obsidian Aide is a lightweight, distraction-free AI companion designed for deep thinking, note-taking, and writing within Obsidian. Unlike bloated developer-heavy plugins, Aide focuses on pure conversational interaction and symbiotic knowledge creation. Users can enjoy built-in web search, tool integration, and multi-provider AI chat out of the box with zero runtime setup required.

---

🚀 Key Features & Highlights

1. Lightweight & Content-Focused Conversation:

   • Zero engineering noise: stripped away intrusive Diff/Apply diff merges and cluttered controls, returning to pure natural reading and writing;
   • Ultra-lightweight distribution: final bundled size is only ~2.1 MB (over 75% reduction compared to traditional alternatives), with zero redundant dependencies and instant startup.

2. Modern Native Chat Experience:

   • High-fidelity SMS-style bubble interface styled with native Obsidian CSS variables, seamlessly adapting to light, dark, and community themes (e.g., Minimal);
   • Foldable callout-style rendering for reasoning/thinking models (e.g., DeepSeek-R1, Qwen-Thinking) with smooth streaming expansion;
   • In-place editing of historical questions with instant regeneration.

3. Vault Knowledge Copilot:

   • Type @ to instantly search, link, and reference specific documents or folders in your vault;
   • Chat alongside notes: automatically detects current active notes for intelligent summarization, key takeaways, and multi-document synthesis.

4. Free-Tier Friendly & Anti-429 Optimization:

   • Built-in 🌱 Eco Mode (Lightweight Free-Tier) and 🚀 Pro Mode (High-Precision);
   • Deeply optimized for free models on Groq (8,000 TPM limit) and SiliconFlow, reducing per-query tokens by up to 75%;
   • Intelligent 429 rate limit backoff and silent retries with friendly countdown alerts;
   • Proprietary Loop Guard prevents infinite tool-calling loops and guarantees structured conclusions.

5. MCP Friendly & Extensible Tools Ecosystem:

   • Built-in support for the standard Model Context Protocol (MCP), allowing connection to external MCP servers (via stdio / SSE) to expand AI capabilities without limits;
   • Out-of-the-box native tools without Node.js runtime required:
     • Domestic Bing Search (CN): Direct connection, no proxy, no API key needed;
     • Web Fetch & Parser: Converts any webpage URL into clean, readable Markdown;
     • Real-Time Weather, arXiv Paper Search, and Accurate Local System Time.

---

📥 Installation Guide

1. Download obsidian-aide.zip from the release assets below;
2. Extract the archive to find three files: main.js, manifest.json, and styles.css;
3. Move these three files into your Obsidian vault under .obsidian/plugins/obsidian-aide/;
4. Go to Obsidian Settings -> Community plugins, reload, and enable Aide!

---

🙏 Acknowledgments

This project is inspired by and built upon the excellent open-source project obsidian-smart-composer. Sincere thanks to the original author glowingjade!

---

🇨🇳 中文说明 (Chinese)

🎉 欢迎使用 Obsidian Aide！

Obsidian Aide 是一款轻量级、去干扰、以深度思考与创作为中心的 Obsidian AI 伴侣。不同于臃肿复杂的开发型插件，Aide 回归纯粹的聊天交互与笔记共生体验，普通用户无需安装 Node.js 即可开箱畅享免翻墙联网搜索与多模型问答。

---

🚀 核心特性与亮点 (Highlights)

1. 轻量对话聚焦内容：

   • 去工程化噪音：彻底剔除侵入式的 Diff/Apply 代码合并等繁杂功能与冗余标签，回归纯粹的 Markdown 思考阅读与创作伴侣；
   • 极致轻量架构：最终打包产物体积仅 2.10 MB（较传统同类插件减重 75% 以上），零多余依赖，极速秒开加载。

2. 现代聊天体验：

   • 短信式原生对仗对话体验：用户提问气泡与 AI 回复卡片高保真绑定 Obsidian 官方设计变量，自适应浅色、深色及第三方主题（如 Minimal）；
   • 深度思考优雅折叠：支持推理模型（如 DeepSeek-R1、Qwen-Thinking）的思考过程类 Callout 式折叠与流式展开，思绪清晰可溯；
   • 就地修改历史提问并重新生成。

3. 知识库功能：

   • 文件与文件夹精准引用：输入 @ 即可智能检索并关联当前笔记库内的特定文档或目录；
   • 当前文档边读边聊：自动感知当前编辑焦点，针对长笔记进行智能总结、提炼要点与多方对比。

4. 免费层友好：

   • 内置“🌱 免费层轻量模式 (Eco)”与“🚀 付费层高精度模式 (Pro)”自由切换；
   • 深度适配 Groq（8,000 TPM 限额）与 SiliconFlow 免费模型，单次提问 Token 骤降 75%；
   • 智能捕获 429 速率限制并自动静默重试，告别突兀大红屏；
   • 独创循环熔断机制（Loop Guard），彻底根治模型在工具调用中无休止死磕的死循环。

5. MCP 友好（工具扩展生态）：

   • 深度集成开放的标准 Model Context Protocol (MCP) 架构，支持连接外部 MCP 服务器（stdio / SSE），无限扩充 AI 工具能力；
   • 免 Node.js 原生工具开箱即用：
     • 国内必应联网搜索 (Bing CN)：直连国内服务器，免翻墙、免 API Key；
     • 网页正文深度提取 (Web Fetch)：自动解析任意网页为干净的 Markdown 正文；
     • 实时天气预报、arXiv 前沿学术论文检索、精准当前时间随叫随到。

---

📥 安装使用指南

1. 下载下方附件中的 obsidian-aide.zip；
2. 解压得到三个文件：main.js、manifest.json、styles.css；
3. 将这三个文件放置于您的 Obsidian 库目录中的 .obsidian/plugins/obsidian-aide/ 文件夹下；
4. 进入 Obsidian 【设置】->【第三方插件】，点击刷新并开启 Aide 即可！

---

🙏 致谢

本项目在产品设计理念与底层架构上深受开源项目 obsidian-smart-composer 的启发，在此对原作者 glowingjade 表达诚挚的感谢！

0.8.3

Release Notes: v0.8.3 - Obsidian Aider Official Release / 正式发布

🌐 English

🎉 Welcome to Obsidian Aider!

Obsidian Aider is a lightweight, distraction-free AI companion designed for deep thinking, note-taking, and writing within Obsidian. Unlike bloated developer-heavy plugins, Aider focuses on pure conversational interaction and symbiotic knowledge creation. Users can enjoy built-in web search, tool integration, and multi-provider AI chat out of the box with zero runtime setup required.

---

🚀 Key Features & Highlights

1. Lightweight & Content-Focused Conversation:

   • Zero engineering noise: stripped away intrusive Diff/Apply diff merges and cluttered controls, returning to pure natural reading and writing;
   • Ultra-lightweight distribution: final bundled size is only ~2.1 MB (over 75% reduction compared to traditional alternatives), with zero redundant dependencies and instant startup.

2. Modern Native Chat Experience:

   • High-fidelity SMS-style bubble interface styled with native Obsidian CSS variables, seamlessly adapting to light, dark, and community themes (e.g., Minimal);
   • Foldable callout-style rendering for reasoning/thinking models (e.g., DeepSeek-R1, Qwen-Thinking) with smooth streaming expansion;
   • In-place editing of historical questions with instant regeneration.

3. Vault Knowledge Copilot:

   • Type @ to instantly search, link, and reference specific documents or folders in your vault;
   • Chat alongside notes: automatically detects current active notes for intelligent summarization, key takeaways, and multi-document synthesis.

4. Free-Tier Friendly & Anti-429 Optimization:

   • Built-in 🌱 Eco Mode (Lightweight Free-Tier) and 🚀 Pro Mode (High-Precision);
   • Deeply optimized for free models on Groq (8,000 TPM limit) and SiliconFlow, reducing per-query tokens by up to 75%;
   • Intelligent 429 rate limit backoff and silent retries with friendly countdown alerts;
   • Proprietary Loop Guard prevents infinite tool-calling loops and guarantees structured conclusions.

5. MCP Friendly & Extensible Tools Ecosystem:

   • Built-in support for the standard Model Context Protocol (MCP), allowing connection to external MCP servers (via stdio / SSE) to expand AI capabilities without limits;
   • Out-of-the-box native tools without Node.js runtime required:
     • Domestic Bing Search (CN): Direct connection, no proxy, no API key needed;
     • Web Fetch & Parser: Converts any webpage URL into clean, readable Markdown;
     • Real-Time Weather, arXiv Paper Search, and Accurate Local System Time.

---

📥 Installation Guide

1. Download obsidian-aider.zip (or aider.zip) from the release assets below;
2. Extract the archive to find three files: main.js, manifest.json, and styles.css;
3. Move these three files into your Obsidian vault under .obsidian/plugins/aider/;
4. Go to Obsidian Settings -> Community plugins, reload, and enable Aider!

---

🙏 Acknowledgments

This project is inspired by and built upon the excellent open-source project obsidian-smart-composer. Sincere thanks to the original author glowingjade!

---

🇨🇳 中文说明 (Chinese)

🎉 欢迎使用 Obsidian Aider！

Obsidian Aider 是一款轻量级、去干扰、以深度思考与创作为中心的 Obsidian AI 伴侣。不同于臃肿复杂的开发型插件，Aider 回归纯粹的聊天交互与笔记共生体验，普通用户无需安装 Node.js 即可开箱畅享免翻墙联网搜索与多模型问答。

---

🚀 核心特性与亮点 (Highlights)

1. 轻量对话聚焦内容：

   • 去工程化噪音：彻底剔除侵入式的 Diff/Apply 代码合并等繁杂功能与冗余标签，回归纯粹的 Markdown 思考阅读与创作伴侣；
   • 极致轻量架构：最终打包产物体积仅 2.10 MB（较传统同类插件减重 75% 以上），零多余依赖，极速秒开加载。

2. 现代聊天体验：

   • 短信式原生对仗对话体验：用户提问气泡与 AI 回复卡片高保真绑定 Obsidian 官方设计变量，自适应浅色、深色及第三方主题（如 Minimal）；
   • 深度思考优雅折叠：支持推理模型（如 DeepSeek-R1、Qwen-Thinking）的思考过程类 Callout 式折叠与流式展开，思绪清晰可溯；
   • 就地修改历史提问并重新生成。

3. 知识库功能：

   • 文件与文件夹精准引用：输入 @ 即可智能检索并关联当前笔记库内的特定文档或目录；
   • 当前文档边读边聊：自动感知当前编辑焦点，针对长笔记进行智能总结、提炼要点与多方对比。

4. 免费层友好：

   • 内置“🌱 免费层轻量模式 (Eco)”与“🚀 付费层高精度模式 (Pro)”自由切换；
   • 深度适配 Groq（8,000 TPM 限额）与 SiliconFlow 免费模型，单次提问 Token 骤降 75%；
   • 智能捕获 429 速率限制并自动静默重试，告别突兀大红屏；
   • 独创循环熔断机制（Loop Guard），彻底根治模型在工具调用中无休止死磕的死循环。

5. MCP 友好（工具扩展生态）：

   • 深度集成开放的标准 Model Context Protocol (MCP) 架构，支持连接外部 MCP 服务器（stdio / SSE），无限扩充 AI 工具能力；
   • 免 Node.js 原生工具开箱即用：
     • 国内必应联网搜索 (Bing CN)：直连国内服务器，免翻墙、免 API Key；
     • 网页正文深度提取 (Web Fetch)：自动解析任意网页为干净的 Markdown 正文；
     • 实时天气预报、arXiv 前沿学术论文检索、精准当前时间随叫随到。

---

📥 安装使用指南

1. 下载下方附件中的 obsidian-aider.zip（或 aider.zip）；
2. 解压得到三个文件：main.js、manifest.json、styles.css；
3. 将这三个文件放置于您的 Obsidian 库目录中的 .obsidian/plugins/aider/ 文件夹下；
4. 进入 Obsidian 【设置】->【第三方插件】，点击刷新并开启 Aider 即可！

---

🙏 致谢

本项目在产品设计理念与底层架构上深受开源项目 obsidian-smart-composer 的启发，在此对原作者 glowingjade 表达诚挚的感谢！