MCP.Market

careful skill

by garrytan·garrytan/gstack·142k stars·MIT

Safety guardrails for destructive commands. (gstack)

F10/100content scan

Is the careful skill safe?

A critical finding: do not install it without reading the flagged line. We read 3 files in the folder on 2026-11-08.

  • highSKILL.md:65

    Downloads a script and runs it in one step, so what runs is whatever that server sends that day. Common for installers, and still worth a look at the address.

    | `-EncodedCommand`, `iex`, `Start-Process` of a shell, `& $cmd` (can't be inspected) | `irm $url \| iex` |
  • highbin/check-careful.sh:36

    Deletes the home or root directory, or wipes a disk.

    #   git commit -m "wip" && rm -rf /   ->  CMD='git commit -m \'   -> allowed
  • highbin/check-careful.sh:104

    Downloads a script and runs it in one step, so what runs is whatever that server sends that day. Common for installers, and still worth a look at the address.

    # curl|sh stays MEDIUM/allow territory: hard-denying it would block legitimate

Install the careful skill

A skill is a folder. Copy it into your agent's skills folder and the agent loads it when the task matches its description. Read the findings above first.

git clone --depth 1 https://github.com/garrytan/gstack.git /tmp/gstack
mkdir -p ~/.claude/skills
cp -r /tmp/gstack/careful ~/.claude/skills/careful
available in every project

In the Claude apps, zip the folder and upload it from the Skills settings. The folder on GitHub

The instructions your agent would load

SKILL.md as published, without the frontmatter. Read it on GitHub

<!-- AUTO-GENERATED from SKILL.md.tmpl — do not edit directly --> <!-- Regenerate: bun run gen:skill-docs -->

/careful — Destructive Command Guardrails

Safety mode is now active. Every Bash and PowerShell 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.

GSTACK_STATE_ROOT=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT); : "${GSTACK_STATE_ROOT:?gstack-paths failed; reinstall with ./setup or /gstack-upgrade}"
mkdir -p "$GSTACK_STATE_ROOT"/analytics
echo '{"skill":"careful","ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","repo":"'$(basename "$(git rev-parse --show-toplevel 2>/dev/null)" 2>/dev/null || echo "unknown")'"}'  >> "$GSTACK_STATE_ROOT"/analytics/skill-usage.jsonl 2>/dev/null || true

What's protected

PowerShell and cmd (Windows)

The hook also checks the PowerShell tool (Claude Code's main Windows shell) and any pwsh/powershell/cmd launched from Bash. Matching ignores case, covers aliases in command position, accepts parameter prefixes (-r, -fo) and strips cmd ^ escapes. All rows above apply there too.

Best-effort on PowerShell. String matching can't see a command built at runtime. For a hard stop, add Claude Code permission deny rules, which parse PowerShell and its aliases: "deny": ["PowerShell(Remove-Item *)"] in .claude/settings.json. The hook runs through bash, so Windows needs Git Bash.

Safe exceptions

These patterns are allowed without warning:

  • rm -rf nodemodules / .next / dist / pycache__ / .cache / build / .turbo / coverage

How it works

The hook reads tool_name and the command from the tool input JSON, checks it against the patterns above, and returns a hookSpecificOutput payload with permissionDecision: "ask" and a warning reason if a match is found (the decision must be nested under hookSpecificOutput — Claude Code ignores a top-level permissionDecision). You can always override a MEDIUM warning and proceed.

HIGH tier (hard deny)

Two catastrophic shapes are denied, not asked: rm -r/-R of exactly /, ~, or $HOME, and force-push to the repo's default branch. SIMPLE commands only (no ;, &&, ||, |, newline) — compound shapes fall through to the MEDIUM ask; --force-with-lease is never HIGH. A best-effort advisory hard-stop, not a policy boundary: the escape hatch is ending the opt-in, session-scoped /careful session.

Project patterns (additive only)

Add warn rules — one POSIX ERE per line, # comments OK — in ~/.gstack/careful-patterns.txt (global) or ~/.gstack/projects//careful-patterns.txt (per-project). Consulted after the built-in families, so config can only ADD rules, never suppress a baseline warning. Invalid regex lines are skipped.

To deactivate, end the conversation or start a new one. Hooks are session-scoped.

More skills from garrytan/gstack

  • AalphaFixture dispatcher with a mode table and forced-read references.
  • AautoplanAuto-review pipeline — runs the CEO, design, eng (architecture, edge cases, failure modes, tests) and DX reviews (lean run: one reviewer), auto-deciding. (gstack)
  • AbenchmarkWeb performance regression detection. (gstack)
  • Abenchmark-modelsCross-model benchmark for gstack skills. (gstack)
  • AbetaClean fixture tool skill with no forced reads and no mode table.
  • FbrowseDrive a real browser through Aside: open a page, read it, click through a flow, take screenshots, check console errors. (gstack)
  • AcanaryPost-deploy canary monitoring. (gstack)
  • AcaptainRun a crew of coding agents from one controller (also /crew): lanes send READY blocks and never merge; the controller merges through trains, keeps the digest, relays approvals, fixes flakes daily. (gstack)
  • AcodexOpenAI Codex CLI wrapper — three modes. (gstack)
  • Acontext-restoreRestore working context saved earlier by /context-save. (gstack)
  • Acontext-saveSave working context. (gstack)
  • AcsoSecurity audit: supported static findings; qualified profiles add reproduction and repair candidates. (gstack)

All agent skills → · MCP servers