safety-guard
三種模式防止破壞性操作：危險指令確認、鎖定編輯目錄、兩者合併
中文翻譯
--- name: safety-guard description: "用三種模式防範破壞性操作：Careful 攔下危險指令（rm -rf、git push --force、DROP TABLE）要求確認，Freeze 把寫入鎖在單一目錄，Guard 兩者合併。這只是建議性質：由 agent 自己套用這些規則，沒有附帶 hook。在正式環境作業、讓 agent 自主執行、限制只能編輯某個目錄，或進行 migration、deploy、資料變更時使用。" metadata: origin: ECC --- # Safety Guard：防止破壞性操作 ## 何時使用 - 在正式環境作業時 - Agent 自主執行時（full-auto 模式） - 想把編輯限制在特定目錄時 - 進行敏感操作時（migration、deploy、資料變更） ## 運作方式 三種保護模式。它們是 agent 要遵守的指示，不是強制攔截（見[強制執行](#強制執行)）。 ### 模式 1：Careful 模式 執行破壞性指令之前，agent 會先停下來警告： ``` Watched patterns: - rm -rf (especially /, ~, or project root) - git push --force - git reset --hard - git checkout . (discard all changes) - DROP TABLE / DROP DATABASE - docker system prune - kubectl delete - chmod 777 - sudo rm - npm publish (accidental publishes) - Any command with --no-verify ``` 偵測到時：說明這個指令會做什麼、請求確認、建議比較安全的替代做法。 ### 模式 2：Freeze 模式 把檔案編輯鎖在特定的目錄樹，例如：「freeze edits to `src/components/`」。 Agent 不會在 `src/components/` 之外 Write/Edit，如果某個改動需要動到外面，會解釋原因。想讓 agent 專心處理一個區塊、不碰無關程式碼時很好用。 ### 模式 3：Guard 模式（Careful + Freeze 合併） 兩種保護同時啟動，例如：「guard mode, write only to `src/api/`, read anything」。 Agent 什麼都能讀，但只會寫入 `src/api/`，而且執行任何破壞性指令前都會先問。 ### 解除 告訴 agent 不要再套用這些規則。 ## 強制執行 這個 skill 只是建議性質。ECC 沒有附 `/safety-guard` 指令，也沒有對應的 hook，所以違反這些規則的 tool call 不會被擋下來。想要不靠 agent 的強制攔截，請用 Claude Code 的[權限 deny 規則](https://code.claude.com/docs/en/permissions)（例如 `Bash(rm -rf *)`）或 [PreToolUse hook](https://code.claude.com/docs/en/hooks)。 ## 整合 - `codex -a never` 的 session 預設啟用 - 搭配 ECC 2.0 的可觀測性風險評分一起使用
英文原文
--- name: safety-guard description: "Guard against destructive operations with three modes: Careful intercepts dangerous commands (rm -rf, git push --force, DROP TABLE) for confirmation, Freeze locks writes to one directory, and Guard combines both. Advisory: the agent applies these rules itself; no hook ships with it. Use when working on production systems, running agents autonomously, restricting edits to a directory, or during migrations, deploys, and data changes." metadata: origin: ECC --- # Safety Guard — Prevent Destructive Operations ## When to Use - When working on production systems - When agents are running autonomously (full-auto mode) - When you want to restrict edits to a specific directory - During sensitive operations (migrations, deploys, data changes) ## How It Works Three modes of protection. They are instructions the agent follows, not enforced blocks (see [Enforcement](#enforcement)). ### Mode 1: Careful Mode Before it runs a destructive command, the agent stops and warns: ``` Watched patterns: - rm -rf (especially /, ~, or project root) - git push --force - git reset --hard - git checkout . (discard all changes) - DROP TABLE / DROP DATABASE - docker system prune - kubectl delete - chmod 777 - sudo rm - npm publish (accidental publishes) - Any command with --no-verify ``` When detected: shows what the command does, asks for confirmation, suggests safer alternative. ### Mode 2: Freeze Mode Locks file edits to a specific directory tree, for example: "freeze edits to `src/components/`". The agent does not Write/Edit outside `src/components/`, and explains why when a change would need it. Useful when you want an agent to focus on one area without touching unrelated code. ### Mode 3: Guard Mode (Careful + Freeze combined) Both protections active, for example: "guard mode, write only to `src/api/`, read anything". The agent can read anything but only writes to `src/api/`, and asks before any destructive command. ### Unlock Tell the agent to stop applying the rules. ## Enforcement This skill is advisory. ECC ships no `/safety-guard` command and no hook for it, so nothing blocks a tool call that breaks these rules. For blocking that does not depend on the agent, use Claude Code [permission deny rules](https://code.claude.com/docs/en/permissions) (for example `Bash(rm -rf *)`) or a [PreToolUse hook](https://code.claude.com/docs/en/hooks). ## Integration - Enable by default for `codex -a never` sessions - Pair with observability risk scoring in ECC 2.0