返回工作区根目录、沙箱模式、审批策略和服务状态。

按照 Codex 风格，从工作区根目录到目标目录逐层加载 AGENTS.override.md 或 AGENTS.md。修改代码前应先调用。

读取工作区内文本文件，可指定行范围并返回行号。

一次读取多个工作区文本文件。单次最多 20 个。

列出工作区目录的直接子项。

返回文件或目录的类型、大小、修改时间和权限。

按路径片段和扩展名递归查找文件；自动跳过常见构建目录。

在工作区搜索文本。优先使用 ripgrep，缺失时使用 Node 回退实现。

以原子替换方式创建或完整写入文本文件。已有文件默认允许覆盖；精确修改优先使用 edit_file 或 apply_patch。

仅在 old_text 出现次数与 expected_occurrences 完全一致时替换，避免误改。

应用以 *** Begin Patch / *** End Patch 包裹的 Add、Delete、Update、Move 文件补丁。路径只能相对工作区；所有操作先验证，失败时尝试回滚。

递归创建工作区目录。

在工作区内移动或重命名文件/目录。覆盖目标时需要用户在 interactive_feedback 中批准。

删除工作区内文件或目录。始终通过 interactive_feedback 请求批准。禁止删除工作区根目录。

在工作区执行前台命令。优先传 argv 数组；字符串命令通过系统 shell 执行。高风险命令通过 interactive_feedback 审批。

返回 git status --short --branch。

返回工作树或暂存区差异，可限制路径。

返回最近提交日志。

查看指定提交或对象，可限制文件路径。

维护 Codex 风格任务计划。未全部完成时必须且只能有一个 in_progress；完成后所有步骤为 completed。

返回当前内存中的执行计划。

在最终回答前通过唯一沟通渠道向用户展示工作摘要并等待反馈。timeout 固定为 172800 秒。每次回答结束前都要调用 interactive_feedback 与用户进行交流 这是你与用户之间的唯一沟通渠道。

在工作区或其子目录执行 codegraph init .，建立本地代码知识图谱索引。

在指定目录执行 codegraph sync，同步已有索引。

执行 codegraph status，并返回 MCP 子进程状态。

返回本地工作区、反馈 Web UI 和 CodeGraph 集成状态。

Quick symbol search by name. Returns locations only (no code). Use codegraph_explore instead to get the actual source / understand an area in one call.

List functions that call . For the full flow, use codegraph_explore.

List functions that calls. For the full flow, use codegraph_explore.

List symbols affected by changing . Use before a refactor.

Two modes. (1) READ A FILE — use INSTEAD of the Read tool: pass file (a path or basename) with no symbol and it returns that file's current on-disk source with line numbers, exactly the shape Read gives you (<n>\t<line>, safe to Edit from), narrowable with offset/limit just like Read — PLUS a one-line note of which files depend on it. Same bytes as Read, faster (served from the index), with the blast radius attached. Use it whenever you would Read a source file. (2) ONE SYMBOL you can name — its location, signature, verbatim source (includeCode=true) and caller/callee trail in one call, so before changing it you see what calls it and what your edit would break. For an AMBIGUOUS name it returns EVERY matching definition's body in one call (so you never Read a file to find the right overload); pass file/line to pin one. Use codegraph_explore for several related symbols or the full flow.

PRIMARY TOOL — call FIRST for almost any question OR before an edit: how does X work, architecture, a bug, where/what is X, surveying an area, or the symbols you are about to change. Returns the verbatim source of the relevant symbols grouped by file in ONE capped call (Read-equivalent — treat the shown source as already Read; do NOT re-open those files), plus the call path among them. Query can be a natural-language question OR a bag of symbol/file names. Usually the ONLY call you need — more accurate context, in far fewer tokens and round-trips than a search/Read/Grep loop.

Index health check (files / nodes / edges). Skip unless debugging.

Indexed file tree with language + symbol counts. Faster than Glob for project layout.