Menu
On-demand safety gate that blocks destructive commands. Activate with /learn careful when working on production systems or sensitive data.
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
Locks file editing to the current working directory only. Prevents accidental edits to files outside the project scope.
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
Build MCP servers with Node/TypeScript SDK — tools, resources, prompts, Zod validation, stdio vs Streamable HTTP. Use Context7 or official MCP docs for latest API.
Reference guide for writing Claude Code skills that agents actually follow. Uses 7 persuasion principles to maximize compliance.
Gate that blocks success claims without fresh evidence. Enforces run-the-command-show-the-output before marking work done.
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.
| name | careful |
| description | On-demand safety gate that blocks destructive commands. Activate with /learn careful when working on production systems or sensitive data. |
When this skill is loaded, treat ALL of the following as BLOCKED unless the user explicitly confirms:
rm -rf or rm -r on any directoryDROP TABLE, DROP DATABASE, TRUNCATEgit push --force, git push -fgit reset --harddocker rm, docker rmi, docker system prunekubectl delete--no-verify or --skip-hooksWhen you encounter a blocked operation:
This skill stays active until the session ends or the user says "disable careful mode".