@skyf0xx/hedgehog 6.0.2 → 6.0.4

package/README.md

@@ -95,20 +95,18 @@ Agents fan out to give you great outcomes at **faster speeds**.
 
 ## Your Code is a Graph
 
-Hedgehog indexes your codebase with [CodeGraphContext](https://github.com/CodeGraphContext/CodeGraphContext).
+Hedgehog reaches for your editor's own Language Server Protocol integration to see what depends on what.
 
 ```mermaid
 flowchart TD
-    A[Code Repository] --> B[Tree-sitter / SCIP Indexing]
-    B --> C[Knowledge Graph]
-    C --> D[Graph Database]
-    D --> E[MCP Server]
-    E --> F[Hedgehog]
+    A[Code Repository] --> B[Language Server]
+    B --> C[Symbols & References]
+    C --> D[Hedgehog]
 ```
 
-- **Fewer tokens burned**: tasks arrive knowing what to read, instead of searching for it
-- **Faster builds**: no time spent hunting through the codebase for context
-- **No more surprise breakage**: every task knows what depends on it before it edits
+- **No index to go stale**: the language server's view of the code is always current
+- **No time spent hunting through the codebase for context**: find references and callers directly
+- **No more surprise breakage**: every task can check what depends on it before it edits
 - **Tests cover what changed**: verification that misses affected code gets caught
 
 ## Deterministic Code Generation
@@ -202,9 +200,7 @@ Where possible, Hedgehog uses a battle-tested blueprint in [`hedgehog-core-autho
 
 ### Existing codebases
 
-Hedgehog also adopts onto existing repos.
-
-It scans the repo's shape and is able to create new changes with the same scoped, verified, committed loop.
+Hedgehog also adopts onto a repo it didn't build. It reads the repo and proposes a layer chain that verifies with the repo's own test/lint/build commands.
 
 ## Why Hedgehog Works
 
@@ -220,10 +216,7 @@ Mention `Hedgehog` whenever you want to build something with it.
 <summary>Instructions for AI Agents</summary>
 
 1. Check Node >=22.5.0 is installed.
-2. Check Python >=3.10 is installed — it runs [CodeGraphContext](https://github.com/CodeGraphContext/CodeGraphContext), the code index every project uses.
-3. Check CodeGraphContext is installed and configured.
-4. If anything is missing, offer to set it up with the `hedgehog-code-intelligence-setup` skill — `init` stops without it.
-5. Run the install commands above.
+2. Run the install commands below.
 
 </details>
 
@@ -290,8 +283,6 @@ new one.
 
 - Animation skills vendored from [gsap-skills](https://github.com/greensock/gsap-skills).
 
-- Hedgehog indexes your codebase with [CodeGraphContext](https://github.com/CodeGraphContext/CodeGraphContext).
-
 ## Support Hedgehog
 
 If Hedgehog helps you build better software with AI, **give it a ⭐ on GitHub**.

package/README.zh-CN.md

@@ -95,20 +95,18 @@ npx @skyf0xx/hedgehog graph # 显示构建图
 
 ## 你的代码是一张图
 
-Hedgehog 使用 [CodeGraphContext](https://github.com/CodeGraphContext/CodeGraphContext) 为你的代码库建立索引。
+Hedgehog 借助你编辑器自带的 Language Server Protocol 集成来查看代码间的依赖关系。
 
 ```mermaid
 flowchart TD
-    A[代码仓库] --> B[Tree-sitter / SCIP 索引]
-    B --> C[知识图谱]
-    C --> D[图数据库]
-    D --> E[MCP 服务器]
-    E --> F[Hedgehog]
+    A[代码仓库] --> B[Language Server]
+    B --> C[符号与引用]
+    C --> D[Hedgehog]
 ```
 
-- **更省 Token**：任务开始时就已知道该读什么，无需四处搜索
-- **构建更快**：不必在代码库中翻找上下文
-- **不再意外改坏**：每个任务在动手前就知道有哪些代码依赖它
+- **索引永不过期**：语言服务器看到的代码状态始终是最新的
+- **不必在代码库中翻找上下文**：直接查找引用和调用方
+- **不再意外改坏**：每个任务在动手前都能查一查有哪些代码依赖它
 - **测试覆盖改动**：遗漏受影响代码的验证会被发现
 
 ## 确定性代码生成
@@ -204,7 +202,7 @@ CLI、库、浏览器扩展、数据管道等，都会拥有各自的构建顺
 
 Hedgehog 也可以接入已有仓库。
 
-它会扫描仓库的既有结构，并以同样受限、可验证、可提交的循环方式在其上创建新的改动。
+Hedgehog 也可以接入它并未参与构建的仓库。它会读取该仓库，并提出一条构建层次链，用仓库自身的 test/lint/build 命令来验证。
 
 ## Hedgehog 为什么有效
 
@@ -220,10 +218,7 @@ Hedgehog 也可以接入已有仓库。
 <summary>给 AI Agent 的操作说明</summary>
 
 1. 检查已安装 Node >=22.5.0。
-2. 检查已安装 Python >=3.10 —— 它负责运行 [CodeGraphContext](https://github.com/CodeGraphContext/CodeGraphContext)，即每个项目都会用到的代码索引。
-3. 检查 CodeGraphContext 已安装并完成配置。
-4. 如有缺失，主动提出使用 `hedgehog-code-intelligence-setup` skill 完成配置 —— 否则 `init` 会中止。
-5. 运行上方的安装命令。
+2. 运行上方的安装命令。
 
 </details>