Menu
使用此技能可防止在生产系统上工作或自主运行代理时进行破坏性操作。
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Build a fully automated AI-powered data collection agent for any public source — job boards, prices, news, GitHub, sports, anything. Runs on a schedule, enriches data with a free LLM (Gemini Flash), stores results in Notion/Sheets/Supabase, and learns from user feedback. Runs 100% free on GitHub Actions. Use when the user wants to monitor, collect, or track any public data automatically.
Open plans and HTML artifacts in a local browser canvas where the human annotates elements, chats, and approves or requests changes without leaving the page. Use when presenting a plan for review, or when feedback like "move this, change that" is easier pointed at than typed.
Open plans and HTML artifacts in a local browser canvas where the human annotates elements, chats, and approves or requests changes without leaving the page. Use when presenting a plan for review, or when feedback like "move this, change that" is easier pointed at than typed.
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents. v2.1 adds project-scoped instincts to prevent cross-project contamination.
基于本能的学习系统，通过钩子观察会话，创建带置信度评分的原子本能，并将其进化为技能/命令/代理。v2.1版本增加了项目范围的本能，以防止跨项目污染。
任意の自動コンパクションではなく、タスクフェーズを通じてコンテキストを保持するための論理的な間隔での手動コンパクションを提案します。
| name | safety-guard |
| description | 使用此技能可防止在生产系统上工作或自主运行代理时进行破坏性操作。 |
| origin | ECC |
三种保护模式：
在执行破坏性命令前进行拦截并发出警告：
已监控的模式：
- rm -rf（特别是 /、~ 或项目根目录）
- git push --force
- git reset --hard
- git checkout .（丢弃所有更改）
- DROP TABLE / DROP DATABASE
- docker system prune
- kubectl delete
- chmod 777
- sudo rm
- npm publish（意外发布）
- 任何带有 --no-verify 的命令
检测到时：显示命令功能、请求确认、建议更安全的替代方案。
将文件编辑锁定到特定目录树：
/safety-guard freeze src/components/
任何在 src/components/ 之外的写入/编辑操作都会被阻止并附带说明。适用于希望代理专注于某个区域而不触及无关代码的场景。
双重保护同时生效。为自主代理提供最高安全性。
/safety-guard guard --dir src/api/ --allow-read-all
代理可读取任何内容，但仅能写入 src/api/。破坏性命令在所有位置均被阻止。
/safety-guard off
通过 PreToolUse 钩子拦截 Bash、Write、Edit 和 MultiEdit 工具调用。在执行前根据活动规则检查命令/路径。
codex -a never 会话中启用~/.claude/safety-guard.log