jinzd-ai-cli 0.4.263 → 0.4.265

package/README.md

@@ -1,14 +1,14 @@
 **English** | [中文](README.zh-CN.md)
 
-# ai-cli
-
-<!-- AICLI:DOCS_DEFAULT_WEB_PORT=3000 -->
-<!-- AICLI:DOCS_BUILTIN_TOOL_COUNT=30 -->
-<!-- AICLI:DOCS_REPL_COMMAND_COUNT=48 -->
-<!-- AICLI:DOCS_BUILTIN_PROVIDER_COUNT=10 -->
-<!-- AICLI:DOCS_MEMORY_SOURCE=memory.jsonl -->
-
-> A cross-platform AI coding assistant — CLI, Web UI, and Desktop App — with multi-provider support and agentic tool calling
+# ai-cli
+
+<!-- AICLI:DOCS_DEFAULT_WEB_PORT=3000 -->
+<!-- AICLI:DOCS_BUILTIN_TOOL_COUNT=30 -->
+<!-- AICLI:DOCS_REPL_COMMAND_COUNT=48 -->
+<!-- AICLI:DOCS_BUILTIN_PROVIDER_COUNT=10 -->
+<!-- AICLI:DOCS_MEMORY_SOURCE=memory.jsonl -->
+
+> A cross-platform AI coding assistant — CLI, Web UI, and Desktop App — with multi-provider support and agentic tool calling
 
 [![npm version](https://img.shields.io/npm/v/jinzd-ai-cli)](https://www.npmjs.com/package/jinzd-ai-cli)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
@@ -131,14 +131,14 @@ aicli user delete x          # Delete user
 |----------|--------|-------------|
 | **Claude** | Opus 4, Sonnet 4, Haiku 4 | [console.anthropic.com](https://console.anthropic.com) |
 | **Gemini** | 2.5 Pro, 2.5 Flash | [aistudio.google.com](https://aistudio.google.com) |
-| **DeepSeek** | deepseek-v4-flash (default), deepseek-v4-pro | [platform.deepseek.com](https://platform.deepseek.com) |
+| **DeepSeek** | deepseek-v4-flash (default), deepseek-v4-pro | [platform.deepseek.com](https://platform.deepseek.com) |
 | **OpenAI** | GPT-5.4, GPT-5, GPT-4.1, o3, o4-mini | [platform.openai.com](https://platform.openai.com) |
 | **OpenRouter** | 300+ models (Claude, GPT, Gemini, Llama, Qwen, Mistral...) | [openrouter.ai](https://openrouter.ai) |
 | **Zhipu** | GLM-4, GLM-5 | [open.bigmodel.cn](https://open.bigmodel.cn) |
 | **Kimi** | Kimi K2.6, K2.5, K2 Thinking | [platform.moonshot.cn](https://platform.moonshot.cn) |
-| **Qwen** | qwen-plus, qwen-max, qwen-turbo, qwen-coder-plus | [bailian.console.aliyun.com](https://bailian.console.aliyun.com) |
-| **MiniMax** | MiniMax-M3 (default), M2.7, M2.5, M2.1, M2 | [platform.minimaxi.com](https://platform.minimaxi.com) |
-| **Ollama** | Any locally installed model (Llama, Qwen, Gemma, Mistral...) | No API key — [ollama.com](https://ollama.com) |
+| **Qwen** | qwen-plus, qwen-max, qwen-turbo, qwen-coder-plus | [bailian.console.aliyun.com](https://bailian.console.aliyun.com) |
+| **MiniMax** | MiniMax-M3 (default), M2.7, M2.5, M2.1, M2 | [platform.minimaxi.com](https://platform.minimaxi.com) |
+| **Ollama** | Any locally installed model (Llama, Qwen, Gemma, Mistral...) | No API key — [ollama.com](https://ollama.com) |
 
 Any OpenAI-compatible API can also be used via `customBaseUrls` in config.
 
@@ -172,11 +172,11 @@ AI autonomously invokes these 30 tools during conversations:
 | `edit_file` | write | Precise string replacement with fuzzy matching hints + `replaceAll` mode |
 | `list_dir` | safe | List directory contents |
 | `grep_files` | safe | Regex search across files |
-| `glob_files` | safe | Match files by glob pattern |
-| `web_fetch` | safe | Fetch web pages as Markdown (SSRF-protected) |
-| `web_search` | safe | Keyless Bing/Google web search with weak-result fallback |
-| `google_search` | safe | Google Custom Search API |
-| `run_interactive` | write | Run interactive programs with stdin input; arbitrary executables require confirmation |
+| `glob_files` | safe | Match files by glob pattern |
+| `web_fetch` | safe | Fetch web pages as Markdown (SSRF-protected) |
+| `web_search` | safe | Keyless Bing/Google web search with weak-result fallback |
+| `google_search` | safe | Google Custom Search API |
+| `run_interactive` | write | Run interactive programs with stdin input; arbitrary executables require confirmation |
 | `run_tests` | safe | Auto-detect and run project tests (JUnit XML parsing) |
 | `spawn_agent` | safe | Delegate subtasks to named isolated agents (`agent`: explorer/worker/reviewer/security/tester) |
 | `ask_user` | safe | Pause and ask the user a question |
@@ -193,11 +193,11 @@ AI autonomously invokes these 30 tools during conversations:
 | `notebook_edit` | write | Edit Jupyter notebook cells (add/edit/delete/move) |
 | `find_symbol` | safe | Locate symbol definitions via persistent tree-sitter index (TS/JS/TSX/Python/Go/Rust/Java/C++) |
 | `get_outline` | safe | Enumerate all top-level declarations in one source file |
-| `find_references` | safe | Search indexed files for references to a symbol name |
-| `search_code` | safe | Semantic (meaning-based) code search via local sentence embeddings — bilingual, "grep by meaning" |
-| `recall_memory` | safe | Semantically recall relevant excerpts from earlier chat sessions |
+| `find_references` | safe | Search indexed files for references to a symbol name |
+| `search_code` | safe | Semantic (meaning-based) code search via local sentence embeddings — bilingual, "grep by meaning" |
+| `recall_memory` | safe | Semantically recall relevant excerpts from earlier chat sessions |
 
-**Safety levels**: `safe` = auto-execute, `write` = confirmation required (file-editing tools also show a diff preview), `destructive` = prominent warning + confirmation.
+**Safety levels**: `safe` = auto-execute, `write` = confirmation required (file-editing tools also show a diff preview), `destructive` = prominent warning + confirmation.
 
 ## Key REPL Commands
 
@@ -465,9 +465,9 @@ Environment variables take precedence over config file values:
 | `AICLI_API_KEY_OPENROUTER` | OpenRouter API Key |
 | `AICLI_API_KEY_ZHIPU` | Zhipu API Key |
 | `AICLI_API_KEY_KIMI` | Kimi API Key |
-| `AICLI_API_KEY_QWEN` | Qwen / Alibaba Cloud API Key |
-| `AICLI_API_KEY_MINIMAX` | MiniMax API Key |
-| `AICLI_PROVIDER` | Default provider ID |
+| `AICLI_API_KEY_QWEN` | Qwen / Alibaba Cloud API Key |
+| `AICLI_API_KEY_MINIMAX` | MiniMax API Key |
+| `AICLI_PROVIDER` | Default provider ID |
 | `AICLI_NO_STREAM` | Set to `1` to disable streaming |
 | `HTTPS_PROXY` / `HTTP_PROXY` | Proxy URL |
 
@@ -521,15 +521,15 @@ The Web UI (`aicli web`) provides a full-featured browser interface:
 ## Testing
 
 ```bash
-npm test              # Offline regression suite
+npm test              # Offline regression suite
 npm run test:providers # Real-provider smoke tests (network/API-key dependent)
 npm run test:e2e-web  # Playwright browser smoke; may need browser-process permission in sandboxes
 npm run test:watch    # Watch mode
 ```
 
-Offline tests cover authentication, sessions, tool types and danger levels, permissions, output truncation, diff rendering, edit-file similarity, error hierarchy, config management, env loading, provider registry, web-fetch, grep-files, Hub, dev-state, token estimation, tool registry budgets, parallel tool execution, cost tracking, and session tool history.
+Offline tests cover authentication, sessions, tool types and danger levels, permissions, output truncation, diff rendering, edit-file similarity, error hierarchy, config management, env loading, provider registry, web-fetch, grep-files, Hub, dev-state, token estimation, tool registry budgets, parallel tool execution, cost tracking, and session tool history.
 
-## Releasing (Version Bump → Git Tag → Trusted npm Publish)
+## Releasing (Version Bump → Git Tag → Trusted npm Publish)
 
 > **For contributors and AI agents: the entire release is driven by [`scripts/release.mjs`](scripts/release.mjs). Never manually bump the version or run `npm publish` yourself** — a hand-publish of `0.4.207` shipped a version that reported the wrong number and forced a corrective re-release (`0.4.208`). The script's guards exist precisely to stop that.
 
@@ -537,7 +537,7 @@ Offline tests cover authentication, sessions, tool types and danger levels, perm
 
 - You are on the `main` branch and the changes to ship are already in the working tree.
 - `npx tsc --noEmit` is clean and `npm test` is green (the script enforces both and rolls the bump back on failure).
-- npm Trusted Publishing is configured for `jinzd-ai-cli` and `.github/workflows/release.yml`.
+- npm Trusted Publishing is configured for `jinzd-ai-cli` and `.github/workflows/release.yml`.
 - You have push access to the GitHub remote.
 
 **Step 1 — Write the milestone entry FIRST (Guard A).** Add a one-line entry to the `## 最近里程碑` section of [`CLAUDE.md`](CLAUDE.md) that contains the exact full-width-parenthesis marker `（vX.Y.Z）` for the target version. The script refuses to start without it. (Full narrative goes in `CHANGELOG.md`; this section is just the index line.)
@@ -564,10 +564,32 @@ node scripts/release.mjs patch -m "..." --dry-run
 5. `npm test` must pass (rolls the bump back on failure).
 6. Prints the exact file list entering the release commit and blocks if any looks sensitive (`.env`, `*.pem`, `*.key`, `secret`, …). Override consciously with `--allow-sensitive`.
 7. `git add -A` + commit `"<title> (vX.Y.Z)"` (message passed via argv, not shell-interpolated).
-8. Creates an annotated `vX.Y.Z` release tag.
-9. Pushes `main` and the tag. The pinned GitHub Actions workflow verifies the project and publishes to npm through OIDC, without a long-lived npm token.
-
-**Optional flags:** `--dry-run` · `--skip-tests` · `--no-tag` · `--no-push` · `--allow-sensitive`. The old `--no-publish` flag remains an alias for `--no-tag`.
+8. Creates an annotated `vX.Y.Z` release tag.
+9. Pushes `main` and the tag. The pinned GitHub Actions workflow verifies the project and publishes to npm through OIDC, without a long-lived npm token.
+
+**Optional flags:** `--dry-run` · `--skip-tests` · `--no-tag` · `--no-push` · `--allow-sensitive`. The old `--no-publish` flag remains an alias for `--no-tag`.
+
+### Which files a release touches
+
+Two groups: what **you** write by hand (the script never touches these, and Guard A refuses to start without them) and what the **script** rewrites (never edit these by hand).
+
+| File | Written by | What changes |
+|---|---|---|
+| `CHANGELOG.md` | **By hand, before releasing** | A new `## [X.Y.Z] - YYYY-MM-DD` section with the full narrative: root cause / fix / tests. This is the only place detail lives |
+| `CLAUDE.md` → `## 最近里程碑` | **By hand, before releasing** | **One** index line at the top, which must carry the full-width marker `（vX.Y.Z）` — this exact line is what Guard A checks |
+| `package.json` | Script | `"version"` |
+| `src/core/constants.ts` | Script | The `VERSION` constant (what `aicli --version` and the help banner actually read — drift here ships a package that reports the wrong version) |
+| `CLAUDE.md` → `**当前版本**` | Script | The version number (a *second*, separate spot from the hand-written milestone line) |
+| `docs/USAGE.md`, `docs/USAGE.zh-CN.md`, `docs/ADVANCED*.md`, `docs/TUTORIAL*.md`, `docs/RECIPES*.md`, `SECURITY.md`, `docs/SECURITY*.md` | Script | Only the **document-header** baseline lines (`> **Version**: vX.Y.Z`, security baseline, `当前 vX.Y.Z`). Deliberately first-match-only — "feature introduced in v0.4.246+" markers in the body must keep their original versions |
+| `dist/` | Script | Built once before the test gate so `dist` matches the new version (the `version-drift` guard compares the build output); `prepublishOnly` builds again from scratch at publish time |
+
+**Guard B only watches the first three** (`package.json` / `constants.ts` / `CLAUDE.md` current version) — it re-reads them from disk after the bump and asserts each one landed. Both historical incidents (`0.4.207`, `0.4.239`) were drift across exactly these three, and both happened by bypassing the script.
+
+**If the change itself touches any of the following, sync it too** (guards will catch these, but knowing up front saves a round):
+
+- Adding/removing a provider, built-in tool, or REPL command → the `docs-drift` guard requires the count markers and tables across all four docs to match
+- Adding a provider → it must be added to the matrix in `tests/unit/providers/quirk-contracts.test.ts`
+- Node minimum version, the `bin` name, or recovery-path wording → covered by the `version-drift` guard
 
 ## Documentation
 
@@ -577,10 +599,10 @@ node scripts/release.mjs patch -m "..." --dry-run
 - [`docs/MIGRATION.md`](docs/MIGRATION.md) — Migration guide from Claude Code, Codex, Claude Desktop, and Cursor
 - [`docs/RECIPES.md`](docs/RECIPES.md) — Practical recipes by scenario
 - [`docs/SECURITY.md`](docs/SECURITY.md) — **Security model, deployment checklist, audit history.** Read before exposing `aicli web` to a network.
-- [`CHANGELOG.md`](CHANGELOG.md) — Per-version change log
-- [`CONTRIBUTING.md`](CONTRIBUTING.md) · [`SUPPORT.md`](SUPPORT.md) · [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md) — Community guides
-- [`THIRD_PARTY_NOTICES.md`](THIRD_PARTY_NOTICES.md) — Licenses for vendored browser assets
-- [`docs/OPEN_SOURCE_CHECKLIST.md`](docs/OPEN_SOURCE_CHECKLIST.md) — Repository-owner launch checklist
+- [`CHANGELOG.md`](CHANGELOG.md) — Per-version change log
+- [`CONTRIBUTING.md`](CONTRIBUTING.md) · [`SUPPORT.md`](SUPPORT.md) · [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md) — Community guides
+- [`THIRD_PARTY_NOTICES.md`](THIRD_PARTY_NOTICES.md) — Licenses for vendored browser assets
+- [`docs/OPEN_SOURCE_CHECKLIST.md`](docs/OPEN_SOURCE_CHECKLIST.md) — Repository-owner launch checklist
 - [Chinese README](README.zh-CN.md) — 中文说明文档
 
 ## License

package/README.zh-CN.md

@@ -1,14 +1,14 @@
 [English](README.md) | **中文**
 
-# ai-cli
-
-<!-- AICLI:DOCS_DEFAULT_WEB_PORT=3000 -->
-<!-- AICLI:DOCS_BUILTIN_TOOL_COUNT=30 -->
-<!-- AICLI:DOCS_REPL_COMMAND_COUNT=48 -->
-<!-- AICLI:DOCS_BUILTIN_PROVIDER_COUNT=10 -->
-<!-- AICLI:DOCS_MEMORY_SOURCE=memory.jsonl -->
-
-> 跨平台 AI 编程助手 — CLI 终端、Web 界面、桌面应用三合一，支持 10 大 Provider（含本地 Ollama）与 Agentic 工具调用
+# ai-cli
+
+<!-- AICLI:DOCS_DEFAULT_WEB_PORT=3000 -->
+<!-- AICLI:DOCS_BUILTIN_TOOL_COUNT=30 -->
+<!-- AICLI:DOCS_REPL_COMMAND_COUNT=48 -->
+<!-- AICLI:DOCS_BUILTIN_PROVIDER_COUNT=10 -->
+<!-- AICLI:DOCS_MEMORY_SOURCE=memory.jsonl -->
+
+> 跨平台 AI 编程助手 — CLI 终端、Web 界面、桌面应用三合一，支持 10 大 Provider（含本地 Ollama）与 Agentic 工具调用
 
 [![npm version](https://img.shields.io/npm/v/jinzd-ai-cli)](https://www.npmjs.com/package/jinzd-ai-cli)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
@@ -119,14 +119,14 @@ aicli user delete x          # 删除用户
 |----------|------|-------------|
 | **Claude** | Opus 4, Sonnet 4, Haiku 4 | [console.anthropic.com](https://console.anthropic.com) |
 | **Gemini** | 2.5 Pro, 2.5 Flash | [aistudio.google.com](https://aistudio.google.com) |
-| **DeepSeek** | deepseek-v4-flash（默认）, deepseek-v4-pro | [platform.deepseek.com](https://platform.deepseek.com) |
+| **DeepSeek** | deepseek-v4-flash（默认）, deepseek-v4-pro | [platform.deepseek.com](https://platform.deepseek.com) |
 | **OpenAI** | GPT-5.4, GPT-5, GPT-4.1, o3, o4-mini | [platform.openai.com](https://platform.openai.com) |
 | **OpenRouter** | 300+ 模型（Claude/GPT/Gemini/Llama/Qwen/Mistral...） | [openrouter.ai](https://openrouter.ai) |
-| **Zhipu（智谱）** | GLM-4, GLM-5 | [open.bigmodel.cn](https://open.bigmodel.cn) |
+| **Zhipu（智谱）** | GLM-4, GLM-5 | [open.bigmodel.cn](https://open.bigmodel.cn) |
 | **Kimi** | Kimi K2.6, K2.5, K2 Thinking | [platform.moonshot.cn](https://platform.moonshot.cn) |
-| **Qwen** | qwen-plus, qwen-max, qwen-turbo, qwen-coder-plus | [bailian.console.aliyun.com](https://bailian.console.aliyun.com) |
-| **MiniMax** | MiniMax-M3（默认）、M2.7、M2.5、M2.1、M2 | [platform.minimaxi.com](https://platform.minimaxi.com) |
-| **Ollama** | 任意本地模型（Llama、Qwen、Gemma、Mistral...） | 无需 API Key — [ollama.com](https://ollama.com) |
+| **Qwen** | qwen-plus, qwen-max, qwen-turbo, qwen-coder-plus | [bailian.console.aliyun.com](https://bailian.console.aliyun.com) |
+| **MiniMax** | MiniMax-M3（默认）、M2.7、M2.5、M2.1、M2 | [platform.minimaxi.com](https://platform.minimaxi.com) |
+| **Ollama** | 任意本地模型（Llama、Qwen、Gemma、Mistral...） | 无需 API Key — [ollama.com](https://ollama.com) |
 
 也可通过 `customBaseUrls` 配置接入任意 OpenAI 兼容 API。
 
@@ -160,11 +160,11 @@ AI 在对话中可自主调用 30 个工具：
 | `edit_file` | 写入 | 精确字符串替换，模糊匹配提示 + `replaceAll` 全局替换模式 |
 | `list_dir` | 安全 | 列出目录内容 |
 | `grep_files` | 安全 | 正则搜索文件内容 |
-| `glob_files` | 安全 | 按 glob 模式匹配文件 |
-| `web_fetch` | 安全 | 抓取网页转 Markdown（防 SSRF） |
-| `web_search` | 安全 | 无需 Key 的 Bing/Google 网页搜索，弱结果自动降级 |
-| `google_search` | 安全 | Google 自定义搜索 |
-| `run_interactive` | 写入 | 运行交互式程序并输入；任意可执行程序均需确认 |
+| `glob_files` | 安全 | 按 glob 模式匹配文件 |
+| `web_fetch` | 安全 | 抓取网页转 Markdown（防 SSRF） |
+| `web_search` | 安全 | 无需 Key 的 Bing/Google 网页搜索，弱结果自动降级 |
+| `google_search` | 安全 | Google 自定义搜索 |
+| `run_interactive` | 写入 | 运行交互式程序并输入；任意可执行程序均需确认 |
 | `run_tests` | 安全 | 自动检测并运行测试（JUnit XML 解析） |
 | `spawn_agent` | 安全 | 委派子任务给命名隔离 agent（`agent`: explorer/worker/reviewer/security/tester） |
 | `ask_user` | 安全 | 暂停并向用户提问 |
@@ -181,11 +181,11 @@ AI 在对话中可自主调用 30 个工具：
 | `notebook_edit` | 写入 | 编辑 Jupyter notebook 单元格（增/改/删/移动） |
 | `find_symbol` | 安全 | 按名称精确查找函数/类/方法定义（C1，tree-sitter 索引） |
 | `get_outline` | 安全 | 获取文件的符号大纲（类 → 方法树）（C1） |
-| `find_references` | 安全 | 查找符号的所有引用（C1） |
-| `search_code` | 安全 | 语义代码搜索（C2，自然语言查询，中英文均可） |
-| `recall_memory` | 安全 | 从历史聊天会话中语义召回相关片段 |
+| `find_references` | 安全 | 查找符号的所有引用（C1） |
+| `search_code` | 安全 | 语义代码搜索（C2，自然语言查询，中英文均可） |
+| `recall_memory` | 安全 | 从历史聊天会话中语义召回相关片段 |
 
-**安全级别**：`安全` = 自动执行，`写入` = 需要确认（文件编辑工具还会显示 diff 预览），`破坏性` = 醒目警告 + 确认。
+**安全级别**：`安全` = 自动执行，`写入` = 需要确认（文件编辑工具还会显示 diff 预览），`破坏性` = 醒目警告 + 确认。
 
 ### 文档深入阅读
 
@@ -423,9 +423,9 @@ Auto Mode 不会被项目配置静默开启。用 `/permissions recently-denied`
 | `AICLI_API_KEY_OPENROUTER` | OpenRouter API Key |
 | `AICLI_API_KEY_ZHIPU` | 智谱 API Key |
 | `AICLI_API_KEY_KIMI` | Kimi API Key |
-| `AICLI_API_KEY_QWEN` | Qwen / Alibaba Cloud API Key |
-| `AICLI_API_KEY_MINIMAX` | MiniMax API Key |
-| `AICLI_PROVIDER` | 默认 Provider |
+| `AICLI_API_KEY_QWEN` | Qwen / Alibaba Cloud API Key |
+| `AICLI_API_KEY_MINIMAX` | MiniMax API Key |
+| `AICLI_PROVIDER` | 默认 Provider |
 | `AICLI_NO_STREAM` | 设为 `1` 禁用流式输出 |
 | `HTTPS_PROXY` / `HTTP_PROXY` | 代理地址 |
 
@@ -479,13 +479,13 @@ Web UI（`aicli web`）提供功能完备的浏览器界面：
 ## 测试
 
 ```bash
-npm test              # 运行离线回归测试
+npm test              # 运行离线回归测试
 npm run test:watch    # 监听模式
 ```
 
-离线测试覆盖：认证、会话、工具类型与危险级别、权限、输出截断、diff 渲染、edit-file 相似度、错误层级、配置管理、环境变量、Provider 注册、web-fetch、grep-files、Hub、开发状态、Token 估算、工具注册表预算、并行工具执行、费用追踪和会话工具历史。
+离线测试覆盖：认证、会话、工具类型与危险级别、权限、输出截断、diff 渲染、edit-file 相似度、错误层级、配置管理、环境变量、Provider 注册、web-fetch、grep-files、Hub、开发状态、Token 估算、工具注册表预算、并行工具执行、费用追踪和会话工具历史。
 
-## 发布（版本提升 → Git Tag → npm 可信发布）
+## 发布（版本提升 → Git Tag → npm 可信发布）
 
 > **给贡献者与 AI 模型：整个发布流程一律走 [`scripts/release.mjs`](scripts/release.mjs)，切勿手动 bump 版本号或自己跑 `npm publish`。** 曾经手工 publish 的 `0.4.207` 装上后 `--version` 报错版本号，被迫补发 `0.4.208` 对齐——脚本的四道守卫就是为拦住这类事故而存在。
 
@@ -493,7 +493,7 @@ npm run test:watch    # 监听模式
 
 - 当前在 `main` 分支，待发布的改动已在工作区。
 - `npx tsc --noEmit` 零错误、`npm test` 全绿（脚本会强制校验，失败自动回滚版本号）。
-- 已为 `jinzd-ai-cli` 和 `.github/workflows/release.yml` 配置 npm Trusted Publishing。
+- 已为 `jinzd-ai-cli` 和 `.github/workflows/release.yml` 配置 npm Trusted Publishing。
 - 有 GitHub 远端推送权限。
 
 **第 1 步 —— 先写里程碑条目（Guard A）。** 在 [`CLAUDE.md`](CLAUDE.md) 的 `## 最近里程碑` 一节加一行，条目中必须含目标版本的**全角括号**标记 `（vX.Y.Z）`，否则脚本拒绝启动。（完整叙事写进 `CHANGELOG.md`，本节只留一行式索引。）
@@ -520,17 +520,39 @@ node scripts/release.mjs patch -m "..." --dry-run
 5. `npm test` 必须全绿（失败则回滚版本号）。
 6. 打印进入发布 commit 的完整文件清单，若含疑似敏感文件（`.env`、`*.pem`、`*.key`、`secret`…）直接拦下，确认无误加 `--allow-sensitive`。
 7. `git add -A` + commit `"<标题> (vX.Y.Z)"`（消息走 argv 传参，不经 shell 拼接）。
-8. 创建带注释的 `vX.Y.Z` release tag。
-9. 推送 `main` 和 tag；固定到完整 SHA 的 GitHub Actions 工作流会完成验证，并通过 OIDC 发布到 npm，无需长期 npm token。
-
-**可选 flag：** `--dry-run` · `--skip-tests` · `--no-tag` · `--no-push` · `--allow-sensitive`。旧的 `--no-publish` 仍作为 `--no-tag` 的兼容别名。
-
-## 文档
-
-- [English README](README.md) — English documentation
-- [贡献指南](CONTRIBUTING.md) · [支持指南](SUPPORT.md) · [行为准则](CODE_OF_CONDUCT.md)
-- [第三方许可证声明](THIRD_PARTY_NOTICES.md)
-- [开源前检查清单](docs/OPEN_SOURCE_CHECKLIST.md)
+8. 创建带注释的 `vX.Y.Z` release tag。
+9. 推送 `main` 和 tag；固定到完整 SHA 的 GitHub Actions 工作流会完成验证，并通过 OIDC 发布到 npm，无需长期 npm token。
+
+**可选 flag：** `--dry-run` · `--skip-tests` · `--no-tag` · `--no-push` · `--allow-sensitive`。旧的 `--no-publish` 仍作为 `--no-tag` 的兼容别名。
+
+### 一次发版到底改哪些文件
+
+分两类：**你手写的**（脚本不管，缺了 Guard A 直接拒绝启动）和**脚本自动改的**（不要手动动）。
+
+| 文件 | 谁改 | 改什么 |
+|---|---|---|
+| `CHANGELOG.md` | **手写（发版前）** | 新增 `## [X.Y.Z] - YYYY-MM-DD` 段落，写完整叙事：根因 / 修法 / 测试。这是唯一的详细记录处 |
+| `CLAUDE.md` → `## 最近里程碑` | **手写（发版前）** | 顶部加**一行**索引，必须含全角括号标记 `（vX.Y.Z）`——**Guard A 查的就是这一行** |
+| `package.json` | 脚本 | `"version"` |
+| `src/core/constants.ts` | 脚本 | `VERSION` 常量（`aicli --version` 与 help banner 读的是它，脱同步会发出"装了新版报旧版号"的包） |
+| `CLAUDE.md` → `**当前版本**` | 脚本 | 版本号（与上面手写的里程碑行是两处，别混） |
+| `docs/USAGE.md`、`docs/USAGE.zh-CN.md`、`docs/ADVANCED*.md`、`docs/TUTORIAL*.md`、`docs/RECIPES*.md`、`SECURITY.md`、`docs/SECURITY*.md` | 脚本 | 仅**文档头部**的版本基线行（`> **版本**：vX.Y.Z`、文档安全基线、`当前 vX.Y.Z`）。刻意只替换首个匹配——正文里 `v0.4.246+` 这类"功能引入版本"标记必须保持原值 |
+| `dist/` | 脚本 | 测试前先 build 一次让 dist 与新版本对齐（`version-drift` 守卫拿构建产物比对），publish 时 `prepublishOnly` 再全新构建一次 |
+
+**Guard B 只盯前三处**（`package.json` / `constants.ts` / `CLAUDE.md` 当前版本）——bump 后从磁盘重读逐一断言。历史上 `0.4.207` 和 `0.4.239` 两次事故都是这三处脱同步，且都因为绕开了脚本。
+
+**改动本身若涉及以下内容，别忘了同步**（有守卫会红，但先想到更省一轮）：
+
+- 新增/删除 provider、内置工具、REPL 命令 → `docs-drift` 守卫要求四份文档的计数标记与表格同步
+- 新增 provider → 必须加进 `tests/unit/providers/quirk-contracts.test.ts` 的矩阵
+- Node 版本下限、`bin` 名称、恢复路径文案 → `version-drift` 守卫覆盖
+
+## 文档
+
+- [English README](README.md) — English documentation
+- [贡献指南](CONTRIBUTING.md) · [支持指南](SUPPORT.md) · [行为准则](CODE_OF_CONDUCT.md)
+- [第三方许可证声明](THIRD_PARTY_NOTICES.md)
+- [开源前检查清单](docs/OPEN_SOURCE_CHECKLIST.md)
 
 ## License
 

package/dist/{batch-AN47AYCZ.js → batch-C7SHXRQ2.js}

@@ -1,9 +1,9 @@
 #!/usr/bin/env node
 import {
   ConfigManager
-} from "./chunk-WM4VTGAF.js";
+} from "./chunk-X6NWK2YK.js";
 import "./chunk-TU3L3PW7.js";
-import "./chunk-VKBGOFEO.js";
+import "./chunk-Y7KJPNQI.js";
 import {
   atomicWriteFileSync
 } from "./chunk-IW3Q7AE5.js";

package/dist/{chunk-VT6NRDD2.js → chunk-23SJVOQD.js}

@@ -2,7 +2,7 @@
 import {
   HubAgent,
   isConverged
-} from "./chunk-3ARPAVNV.js";
+} from "./chunk-VOIOUJXI.js";
 
 // src/hub/discuss.ts
 var DiscussionOrchestrator = class {
@@ -21,7 +21,11 @@ var DiscussionOrchestrator = class {
   onRoundReview;
   constructor(config, providers) {
     this.agents = config.roles.map(
-      (role) => new HubAgent(role, providers, config.defaultProvider, config.defaultModel, config.context)
+      (role) => {
+        const agent = new HubAgent(role, providers, config.defaultProvider, config.defaultModel, config.context);
+        agent.thinking = config.thinking;
+        return agent;
+      }
     );
     this.humanSteer = config.humanSteer ?? false;
     this.voteConverge = config.voteConverge ?? false;

package/dist/{chunk-QMB7QYZC.js → chunk-2MR7EP24.js}

@@ -2,7 +2,7 @@
 import {
   CONFIG_DIR_NAME,
   PLUGINS_DIR_NAME
-} from "./chunk-VKBGOFEO.js";
+} from "./chunk-Y7KJPNQI.js";
 import {
   atomicWriteFileSync
 } from "./chunk-IW3Q7AE5.js";

package/dist/chunk-5C76XDDF.js

@@ -0,0 +1,78 @@
+#!/usr/bin/env node
+
+// src/repl/theme.ts
+import chalk from "chalk";
+var DARK_THEME = {
+  prompt: chalk.green,
+  info: chalk.cyan,
+  warning: chalk.yellow,
+  error: chalk.red,
+  success: chalk.green,
+  dim: chalk.dim,
+  accent: chalk.cyan,
+  toolCall: chalk.yellow,
+  toolResult: chalk.green,
+  heading: chalk.bold.cyan
+};
+var LIGHT_THEME = {
+  prompt: chalk.blue,
+  info: chalk.blueBright,
+  warning: chalk.yellow,
+  error: chalk.red,
+  success: chalk.green,
+  dim: chalk.gray,
+  accent: chalk.blueBright,
+  toolCall: chalk.magenta,
+  toolResult: chalk.green,
+  heading: chalk.bold.blue
+};
+function resolveColor(name) {
+  if (name.startsWith("#")) return chalk.hex(name);
+  const parts = name.split(".");
+  let result = chalk;
+  for (const part of parts) {
+    const obj = result;
+    if (obj && typeof obj[part] !== "undefined") {
+      result = obj[part];
+    }
+  }
+  if (typeof result !== "function") {
+    process.stderr.write(`[theme] Warning: unrecognized color "${name}", using default.
+`);
+    return chalk;
+  }
+  return result;
+}
+function buildCustomTheme(base, overrides) {
+  if (!overrides) return base;
+  const result = { ...base };
+  for (const [key, colorName] of Object.entries(overrides)) {
+    if (key in result && colorName) {
+      result[key] = resolveColor(colorName);
+    }
+  }
+  return result;
+}
+var _currentTheme = DARK_THEME;
+function initTheme(themeId = "dark", customColors) {
+  switch (themeId) {
+    case "light":
+      _currentTheme = LIGHT_THEME;
+      break;
+    case "custom":
+      _currentTheme = buildCustomTheme(DARK_THEME, customColors);
+      break;
+    default:
+      _currentTheme = DARK_THEME;
+  }
+}
+var theme = new Proxy(DARK_THEME, {
+  get(_target, prop) {
+    return _currentTheme[prop];
+  }
+});
+
+export {
+  initTheme,
+  theme
+};

package/dist/chunk-6YZEZFX5.js

@@ -0,0 +1,40 @@
+#!/usr/bin/env node
+
+// src/core/chat-params.ts
+function chatParamsFrom(params) {
+  return {
+    temperature: params.temperature,
+    maxTokens: params.maxTokens,
+    timeout: params.timeout,
+    thinking: params.thinking,
+    thinkingEffort: params.thinkingEffort,
+    thinkingBudget: params.thinkingBudget
+  };
+}
+function thinkingParamsFor(allModelParams, model) {
+  const params = allModelParams?.[model];
+  if (!params) return {};
+  const out = {};
+  if (params.thinking !== void 0) out.thinking = params.thinking;
+  if (params.thinkingEffort !== void 0) out.thinkingEffort = params.thinkingEffort;
+  if (params.thinkingBudget !== void 0) out.thinkingBudget = params.thinkingBudget;
+  return out;
+}
+function applyThinkingFlag(fromConfig, flag) {
+  if (flag === void 0) return fromConfig;
+  const mode = flag.toLowerCase();
+  if (mode === "off") return { ...fromConfig, thinking: false };
+  if (mode === "on") return { ...fromConfig, thinking: true };
+  if (mode === "low" || mode === "high" || mode === "max") {
+    return { ...fromConfig, thinking: true, thinkingEffort: mode };
+  }
+  throw new Error(`Invalid --think value "${flag}" (expected: off | on | low | high | max)`);
+}
+var NO_THINKING = { thinking: false };
+
+export {
+  chatParamsFrom,
+  thinkingParamsFor,
+  applyThinkingFlag,
+  NO_THINKING
+};

package/dist/{chunk-TL5MD5XU.js → chunk-AUEX6RNY.js}

@@ -30,6 +30,16 @@ var BaseProvider = class {
   get supportsContentOnlyTee() {
     return this.reliableContentOnlyTee;
   }
+  /**
+   * 该 provider 是否支持 `reasoning_effort` 档位（low/high/max）。目前只有 DeepSeek V4。
+   *
+   * 2026-08-14 审计 P2-03：提示符标签此前无条件渲染成 `[THINK:high]`，于是 Claude
+   * （用 thinkingBudget 计 token 预算）和 GLM（根本没有 effort 概念）也显示一个
+   * 既不生效、也无从设置的档位——屏幕上的假信息比没有信息更坏。
+   */
+  get supportsReasoningEffort() {
+    return false;
+  }
   /**
    * 将 Message[] 转换为 OpenAI API 格式的消息数组。
    * content 为 string 时直接传递；为 MessageContentPart[] 时保留数组格式（vision 请求）。
@@ -1113,10 +1123,12 @@ var OpenAICompatibleProvider = class extends BaseProvider {
       if (!firstChoice) {
         return { content: "", model: response.model, usage: void 0 };
       }
+      const reasoningContent = firstChoice.message.reasoning_content;
       return {
         content: firstChoice.message.content ?? "",
         model: response.model,
-        usage: toUsage(response.usage)
+        usage: toUsage(response.usage),
+        ...reasoningContent ? { reasoningContent } : {}
       };
     } catch (err) {
       throw this.wrapError(err);
@@ -1513,6 +1525,10 @@ var DeepSeekProvider = class extends OpenAICompatibleProvider {
       ...thinkingEnabled ? { reasoning_effort: request.thinkingEffort ?? "high" } : {}
     };
   }
+  /** V4 支持 low/high/max 三档 reasoning_effort（官方 2026-08 更新公告）。 */
+  get supportsReasoningEffort() {
+    return true;
+  }
   /** 禁用流式工具调用，确保 chatWithTools 覆写（代码块检测）生效 */
   enableStreamingToolCalls = false;
   /**

package/dist/{chunk-MTZ6XHJD.js → chunk-CVI3LSGR.js}

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   truncateForPersist
-} from "./chunk-GKI7B4BD.js";
+} from "./chunk-TJBG4F2R.js";
 import {
   APP_NAME,
   CONFIG_DIR_NAME,
@@ -11,7 +11,7 @@ import {
   MCP_PROTOCOL_VERSION,
   MCP_TOOL_PREFIX,
   VERSION
-} from "./chunk-VKBGOFEO.js";
+} from "./chunk-Y7KJPNQI.js";
 import {
   atomicWriteFileSync
 } from "./chunk-IW3Q7AE5.js";

package/dist/{chunk-B27MLWFJ.js → chunk-E3JRGJCJ.js}

@@ -1,10 +1,10 @@
 #!/usr/bin/env node
 import {
   ProviderRegistry
-} from "./chunk-TL5MD5XU.js";
+} from "./chunk-AUEX6RNY.js";
 import {
   ConfigManager
-} from "./chunk-WM4VTGAF.js";
+} from "./chunk-X6NWK2YK.js";
 import {
   getActivePluginAssets,
   getPendingHookTrust,
@@ -14,7 +14,7 @@ import {
   listHooks,
   listInstalledPlugins,
   resetStats
-} from "./chunk-QMB7QYZC.js";
+} from "./chunk-2MR7EP24.js";
 import {
   getGitRoot
 } from "./chunk-HOSJZMQS.js";
@@ -26,7 +26,7 @@ import {
   MCP_PROJECT_CONFIG_NAME,
   MEMORY_FILE_NAME,
   VERSION
-} from "./chunk-VKBGOFEO.js";
+} from "./chunk-Y7KJPNQI.js";
 
 // src/diagnostics/doctor-report.ts
 import { existsSync as existsSync3, statSync as statSync2 } from "fs";

package/dist/{chunk-MJICIIBW.js → chunk-IFTUBTDA.js}

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   VERSION
-} from "./chunk-VKBGOFEO.js";
+} from "./chunk-Y7KJPNQI.js";
 
 // src/cli/ci-format.ts
 var DEFAULT_CI_THRESHOLDS = {

package/dist/{chunk-VUZAKIZ5.js → chunk-LZBD2VGJ.js}

@@ -6,7 +6,7 @@ import {
   MEMORY_FILE_NAME,
   MEMORY_MAX_CHARS,
   MEMORY_STORE_FILE_NAME
-} from "./chunk-VKBGOFEO.js";
+} from "./chunk-Y7KJPNQI.js";
 import {
   DEFAULT_PATTERNS,
   redactString

package/dist/{chunk-3RHY2HF7.js → chunk-M2RKBTVZ.js}

@@ -1,5 +1,5 @@
 // src/core/constants.ts
-var VERSION = "0.4.263";
+var VERSION = "0.4.265";
 var APP_NAME = "ai-cli";
 var CONFIG_DIR_NAME = ".aicli";
 var CONFIG_FILE_NAME = "config.json";

package/dist/{chunk-7YCA2BD3.js → chunk-NJLPA6S4.js}

@@ -1,6 +1,6 @@
 import {
   TEST_TIMEOUT
-} from "./chunk-3RHY2HF7.js";
+} from "./chunk-M2RKBTVZ.js";
 
 // src/tools/builtin/run-tests.ts
 import { execSync, spawnSync } from "child_process";