菜单
Ask for permission before destructive actions and use a concise ACTION/COMMAND/REASON confirmation template. Use when planning deletes, unrecoverable modifications, destructive system changes, git history rewrites/amends, force pushes, or other risky operations.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Retrieves up-to-date documentation, API references, and code examples for any developer technology. Use this skill whenever the user asks about a specific library, framework, SDK, CLI tool, or cloud service -- even for well-known ones like React, Next.js, Prisma, Express, Tailwind, Django, or Spring Boot. Your training data may not reflect recent API changes or version updates. Always use for: API syntax questions, configuration options, version migration issues, "how do I" questions mentioning a library name, debugging that involves library-specific behavior, setup instructions, and CLI tool usage. Use even when you think you know the answer -- do not rely on training data for API details, signatures, or configuration options as they are frequently outdated. Always verify against current docs. Prefer this over web search for library documentation and API details.
Use the MCP grep server (`grep_searchGitHub`) for code search and GitHub code lookups instead of `web_search` or `code_search`. The `code_search` tool triggers an SSH passphrase prompt which blocks the workflow. The MCP grep tool searches literal code patterns across millions of public GitHub repos. Use whenever the user asks to find code examples, search GitHub repos, look up API usage patterns, or understand how libraries are used in real projects.
Teach the user a new skill or concept, within this workspace.
| name | safety-guard |
| description | Ask for permission before destructive actions and use a concise ACTION/COMMAND/REASON confirmation template. Use when planning deletes, unrecoverable modifications, destructive system changes, git history rewrites/amends, force pushes, or other risky operations. |
Use this skill when a planned action may be destructive and the user has not explicitly requested that destructive action.
Do not over-ask. Normal recoverable work in a git-tracked codebase can proceed without confirmation.
Ask for permission before destructive actions when recovery is unclear or unavailable, including:
git commit --amend, rebases, hard resets, and history-filtering commands.git push --force and git push --force-with-lease.If the user explicitly and unambiguously asks for the destructive action, proceed without an extra confirmation.
Use ask_user_question and coalesce related confirmations into as few questions as possible.
The confirmation text must follow this shape:
ACTION: one-line short but understandable description of the action
COMMAND (if applicable): `<command here>`
REASON (if applicable): One-line reason for performing this action
Safety Guard is enforced by the companion Pi extension when installed.
/safety enable turns enforcement on./safety disable turns enforcement off./safety status shows current state.Compatibility aliases are also available:
/permissions enable/permissions disable/permissions status