Add docs directory tree with clickable links and key findings table

- EN README: tree view + summary table with EN doc links - ZH README: tree view + summary table with bilingual links (EN + ZH) - Both tables show topic, key findings, and clickable links per row Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-03 11:34:54 +08:00 · 2026-03-31 19:57:45 +08:00
parent e4fe623ea5
commit 4348930ffe
2 changed files with 56 additions and 16 deletions
--- a/README.md
+++ b/README.md
@@ -7,14 +7,34 @@

 ---

-> **Deep Analysis Reports**: Full source code analysis reports are available in the [`docs/`](docs/) directory (bilingual EN/ZH):
-> - Telemetry & Privacy (opt-out-free data collection pipeline)
-> - Hidden Features & Model Codenames (Tengu, Capybara, Fennec, Numbat)
-> - Undercover Mode (AI attribution stripping in open-source contributions)
-> - Remote Control & Killswitches (managed settings, feature flags)
-> - Future Roadmap (Numbat model, KAIROS autonomous mode, voice input)
->
-> **English**: [`docs/en/`](docs/en/) | **中文**: [`docs/zh/`](docs/zh/)
+## Deep Analysis Reports (`docs/`)
+
+Source code analysis reports derived from decompiled v2.1.88. Bilingual (EN/ZH).
+
+```
+docs/
+├── en/                                    # English
+│   ├── 01-telemetry-and-privacy.md        # Telemetry & Privacy — two-tier analytics pipeline, what's collected, why you can't opt out
+│   ├── 02-hidden-features-and-codenames.md# Hidden Features — model codenames (Capybara/Tengu/Numbat), feature flags, internal vs external user differences
+│   ├── 03-undercover-mode.md              # Undercover Mode — how Anthropic employees hide AI authorship in open-source repos
+│   ├── 04-remote-control-and-killswitches.md # Remote Control — managed settings (accept or die), killswitches, model overrides
+│   └── 05-future-roadmap.md               # Future Roadmap — Numbat (next model), KAIROS autonomous agent, voice mode, unreleased tools
+│
+└── zh/                                    # 中文
+    ├── 01-遥测与隐私分析.md                  # 同上，中文版
+    ├── 02-隐藏功能与模型代号.md
+    ├── 03-卧底模式分析.md
+    ├── 04-远程控制与紧急开关.md
+    └── 05-未来路线图.md
+```
+
+| # | Topic | Key Findings |
+|---|-------|-------------|
+| 01 | **Telemetry & Privacy** | Two analytics sinks (1P → Anthropic, Datadog). Environment fingerprint, process metrics, repo hash on every event. **No UI-exposed opt-out** for 1st-party logging. `OTEL_LOG_TOOL_DETAILS=1` enables full tool input capture. |
+| 02 | **Hidden Features & Codenames** | Animal codenames (Capybara v8, Tengu, Fennec→Opus 4.6, **Numbat** next). Feature flags use random word pairs (`tengu_frond_boric`) to obscure purpose. Internal users get better prompts, verification agents, and effort anchors. Hidden commands: `/btw`, `/stickers`. |
+| 03 | **Undercover Mode** | Anthropic employees auto-enter undercover mode in public repos. Model instructed: *"Do not blow your cover"* — strip all AI attribution, write commits "as a human developer would." **No force-OFF exists.** Raises transparency questions for open-source communities. |
+| 04 | **Remote Control** | Hourly polling of `/api/claude_code/settings`. Dangerous changes show blocking dialog — **reject = app exits**. 6+ killswitches (bypass permissions, fast mode, voice mode, analytics sink). GrowthBook flags can change any user's behavior without consent. |
+| 05 | **Future Roadmap** | **Numbat** codename confirmed. Opus 4.7 / Sonnet 4.8 in development. **KAIROS** = fully autonomous agent mode with `<tick>` heartbeats, push notifications, PR subscriptions. Voice mode (push-to-talk) ready but gated. 17 unreleased tools found. |

 ---

--- a/README_CN.md
+++ b/README_CN.md
@@ -7,14 +7,34 @@

 ---

-> **深度分析文档**: 完整的源码分析报告已发布在 [`docs/`](docs/) 目录下，包含中英双语版本：
-> - 遥测与隐私（无法退出的数据收集管道）
-> - 隐藏功能与模型代号（Tengu、Capybara、Fennec、Numbat）
-> - 卧底模式（在开源项目中隐藏 AI 身份）
-> - 远程控制与紧急开关（托管设置、feature flag）
-> - 未来路线图（Numbat 模型、KAIROS 自主模式、语音输入）
->
-> **英文文档**: [`docs/en/`](docs/en/) | **中文文档**: [`docs/zh/`](docs/zh/)
+## 深度分析文档 (`docs/`)
+
+基于 v2.1.88 反编译源码的分析报告，中英双语。
+
+```
+docs/
+├── en/                                    # English
+│   ├── 01-telemetry-and-privacy.md        # 遥测与隐私 — 双层数据管道，收集了什么，为什么无法退出
+│   ├── 02-hidden-features-and-codenames.md# 隐藏功能 — 模型代号（Capybara/Tengu/Numbat），feature flag，内外用户差异
+│   ├── 03-undercover-mode.md              # 卧底模式 — Anthropic 员工如何在开源项目中隐藏 AI 身份
+│   ├── 04-remote-control-and-killswitches.md # 远程控制 — 托管设置（拒绝就退出），紧急开关，模型覆盖
+│   └── 05-future-roadmap.md               # 未来路线图 — Numbat 模型、KAIROS 自主代理、语音模式、未上线工具
+│
+└── zh/                                    # 中文
+    ├── 01-遥测与隐私分析.md
+    ├── 02-隐藏功能与模型代号.md
+    ├── 03-卧底模式分析.md
+    ├── 04-远程控制与紧急开关.md
+    └── 05-未来路线图.md
+```
+
+| # | 主题 | 核心发现 | 链接 |
+|---|------|---------|------|
+| 01 | **遥测与隐私** | 双层分析管道（1P→Anthropic, Datadog）。环境指纹、进程指标、每个事件携带会话/用户 ID。**没有面向用户的退出开关**。`OTEL_LOG_TOOL_DETAILS=1` 可记录完整工具输入。 | [EN](docs/en/01-telemetry-and-privacy.md) · [中文](docs/zh/01-遥测与隐私分析.md) |
+| 02 | **隐藏功能与代号** | 动物代号体系（Capybara v8, Tengu, Fennec→Opus 4.6, **Numbat** 下一代）。Feature flag 用随机词对掩盖用途。内部用户获得更好的 prompt 和验证代理。隐藏命令：`/btw`、`/stickers`。 | [EN](docs/en/02-hidden-features-and-codenames.md) · [中文](docs/zh/02-隐藏功能与模型代号.md) |
+| 03 | **卧底模式** | Anthropic 员工在公开仓库自动进入卧底模式。模型指令："**不要暴露你的掩护身份**" — 剥离所有 AI 归属，commit 看起来像人类写的。**没有强制关闭选项。** | [EN](docs/en/03-undercover-mode.md) · [中文](docs/zh/03-卧底模式分析.md) |
+| 04 | **远程控制与 Killswitch** | 每小时轮询 `/api/claude_code/settings`。危险变更弹出阻塞对话框 — **拒绝 = 程序退出**。6+ 紧急开关（绕过权限、快速模式、语音模式、分析 sink）。GrowthBook 可无同意改变任何用户行为。 | [EN](docs/en/04-remote-control-and-killswitches.md) · [中文](docs/zh/04-远程控制与紧急开关.md) |
+| 05 | **未来路线图** | **Numbat** 代号确认。Opus 4.7 / Sonnet 4.8 开发中。**KAIROS** = 完全自主代理模式，心跳 `<tick>`、推送通知、PR 订阅。语音模式（push-to-talk）已就绪。发现 17 个未上线工具。 | [EN](docs/en/05-future-roadmap.md) · [中文](docs/zh/05-未来路线图.md) |

 ---