gstack-careful
Skill2 starsChanged 23 days agoApache-2.0View on GitHub
Safety guardrails for destructive commands. (gstack)
- Deletes or force-pushes
- Commits and pushes
2 other projects use a version of this, 2 of them edited.
Adopted by
Most recently taken up by
- robertkarl/rkstackedited, first seen Oct 20, 2026
- sagnikmitra/skillsedited, first seen Jun 8, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 2
- sagnikmitra/skills★ 156% same
- robertkarl/rkstack★ 069% same
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository40 other files it gives its agents.
- caveman-commitSkill
- caveman-reviewSkill
- cavemanSkill
- gstack-autoplanSkill
- gstack-csoSkill
- gstack-design-reviewSkill
- gstack-freezeSkill
- gstack-guardSkill
- gstack-investigateSkill
- gstack-office-hoursSkill
- gstack-plan-ceo-reviewSkill
- gstack-plan-design-reviewSkill
- gstack-plan-devex-reviewSkill
- gstack-plan-eng-reviewSkill
- gstack-qaSkill
- gstack-retroSkill
- gstack-reviewSkill
- gstack-shipSkill
- ponytail-auditSkill
- ponytail-debtSkill
- ponytail-reviewSkill
- ponytailSkill
- brainstormingSkill
- dispatching-parallel-agentsSkill
- executing-plansSkill
- finishing-a-development-branchSkill
- receiving-code-reviewSkill
- requesting-code-reviewSkill
- subagent-driven-developmentSkill
- systematic-debuggingSkill
- test-driven-developmentSkill
- using-git-worktreesSkill
- using-superpowersSkill
- verification-before-completionSkill
- writing-plansSkill
- writing-skillsSkill
- tandem-db-safetySkill
- tandem-harnessSkill
- tandem-reviewSkill
- tandem-worktree-etiquetteSkill
The file
Files
Tandem vendoring note. Vendored from gstack (MIT, (c) Garry Tan, https://github.com/garrytan/gstack @ a3259400). The upstream Claude-Code-specific boilerplate (telemetry, session tracking, update checks, plan-mode hooks, learnings/brain sync) was removed; the workflow itself is unchanged. Paths below are relative to this skill's directory; sibling skills live at
../<id>/(ids are prefixedgstack-).$GSTACK_ROOTis an optional upstream gstack install (default~/.claude/skills/gstack) needed only for optional helpers such as thebrowsebinary; skip any step that needs it when it is absent. "Ask the user" means a plain chat question when no structured question tool exists.
When to invoke this skill
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".
/careful — Destructive Command Guardrails
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.
What's protected
| 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 |
Safe exceptions
These patterns are allowed without warning:
rm -rf node_modules/.next/dist/__pycache__/.cache/build/.turbo/coverage
How it works
The 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.
Enforcement outside Claude Code (Tandem note)
Upstream enforces this skill with a Claude Code PreToolUse hook. Other agents have no such hook, so you must self-enforce: before every shell command or file edit, run the matching check script from this skill's bin/ directory (bin/check-careful.sh / bin/check-freeze.sh; they read the tool-input JSON on stdin) or apply the same rules by hand, and stop to ask the user whenever a rule trips. Tandem's own sandbox and permission cards remain the hard boundary.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.