skills/gstack-careful/SKILL.md
Safety guardrails for destructive commands. Warns before rm -rf, DROP TABLE, force-push, git reset --hard, kubectl delete, and similar destructive operations. User can override each warning. Use when touching prod, debugging live systems, or working in a shared environment. Use when asked to "be careful", "safety mode", "prod mode", or "careful mode".
npx skillsauth add Andy-HNU/AndyClaw carefulInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Safety mode is now active. Every bash command will be checked for destructive patterns before running. If a destructive command is detected, you'll be warned and can choose to proceed or cancel.
mkdir -p ~/.gstack/analytics
echo '{"skill":"careful","ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","repo":"'$(basename "$(git rev-parse --show-toplevel 2>/dev/null)" 2>/dev/null || echo "unknown")'"}' >> ~/.gstack/analytics/skill-usage.jsonl 2>/dev/null || true
| Pattern | Example | Risk |
|---------|---------|------|
| rm -rf / rm -r / rm --recursive | rm -rf /var/data | Recursive delete |
| DROP TABLE / DROP DATABASE | DROP TABLE users; | Data loss |
| TRUNCATE | TRUNCATE orders; | Data loss |
| git push --force / -f | git push -f origin main | History rewrite |
| git reset --hard | git reset --hard HEAD~3 | Uncommitted work loss |
| git checkout . / git restore . | git checkout . | Uncommitted work loss |
| kubectl delete | kubectl delete pod | Production impact |
| docker rm -f / docker system prune | docker system prune -a | Container/image loss |
These patterns are allowed without warning:
rm -rf node_modules / .next / dist / __pycache__ / .cache / build / .turbo / coverageThe hook reads the command from the tool input JSON, checks it against the
patterns above, and returns permissionDecision: "ask" with a warning message
if a match is found. You can always override the warning and proceed.
To deactivate, end the conversation or start a new one. Hooks are session-scoped.
tools
Compact long Telegram chat sessions and prevent silent non-replies caused by high context usage. Use when a Telegram conversation becomes slow, misses replies, or session tokens are high (for example above ~80-85%), and when the user asks to clean, rotate, or shrink Telegram context.
tools
Send local files to Telegram chats through OpenClaw CLI delivery. Use when a user asks to send or forward a file (md/txt/pdf/image/video) to Telegram, especially from workspace paths. Supports direct chat targets and group/topic targets.
tools
轻量自我改进记录技能。用于把错误、用户纠正、知识缺口、功能诉求写入 `.learnings/`,并在每天回顾时将高价值经验提升到 AGENTS.md / SOUL.md / TOOLS.md / MEMORY.md。适用于“避免重复犯错、沉淀经验、降低长期 token 浪费”的场景;默认不启用高频 hooks。
data-ai
轻量主动代理增强。用于在不显著增加 token 的前提下引入 4 个高价值机制:VBR(先验证再汇报)、Autonomous vs Prompted Cron 区分、工具迁移检查清单、会话压缩恢复流程。适用于避免“改了文本没改机制”、减少自动任务失效、提高会话连续性。