agentleFS
manastalukdarmanastalukdar/ai-devstudio/skills/git-guardrails/SKILL.md

git-guardrails

Skill1 starsChanged 7 months agoMITView on GitHub

Install a Claude Code PreToolUse hook that blocks dangerous git operations (force push, reset --hard, clean -f, branch -D, checkout ., restore .) before they execute. Use when setting up git safety hooks or preventing destructive git commands in Claude Code.

  • Deletes or force-pushes
  • Commits and pushes

9 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use zoom-out and code-review.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • victorgarciaesgiSkill

    zoom-out

    Tell the agent to zoom out and give broader context or a higher-level perspective. Use when you're unfamiliar with a section of code or need to understand how it fits into the bigger picture.

    Also in 3 projects that use this

  • GoogleSkill

    code-review

    Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.

    Also in 4 projects that use this

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read or create the settings file, then add the hook:

    line 70

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Git Guardrails

Install a PreToolUse hook in Claude Code that intercepts and blocks dangerous git operations before they run.

Arguments: $ARGUMENTS - global to install in ~/.claude/settings.json, or blank for project-level .claude/settings.json

Behavior

1. Determine Scope

if [[ "$ARGUMENTS" == "global" ]]; then
  SETTINGS="$HOME/.claude/settings.json"
  HOOKS_DIR="$HOME/.claude/hooks"
else
  SETTINGS=".claude/settings.json"
  HOOKS_DIR=".claude/hooks"
fi
mkdir -p "$HOOKS_DIR"

2. Create the Block Script

cat > "$HOOKS_DIR/block-dangerous-git.sh" << 'SCRIPT'
#!/bin/bash
set -e

input=$(cat)
command=$(echo "$input" | jq -r '.command // empty')

# Only check Bash tool calls
[ -z "$command" ] && exit 0

dangerous_patterns=(
  "git push"
  "push --force"
  "git reset --hard"
  "reset --hard"
  "git clean -fd"
  "git clean -f"
  "git branch -D"
  "git checkout \."
  "git restore \."
)

for pattern in "${dangerous_patterns[@]}"; do
  if echo "$command" | grep -E "$pattern" > /dev/null 2>&1; then
    echo "BLOCKED: '$command' matches dangerous pattern '$pattern'. The user has prevented this operation." >&2
    exit 2
  fi
done

exit 0
SCRIPT

chmod +x "$HOOKS_DIR/block-dangerous-git.sh"
echo "Hook script created: $HOOKS_DIR/block-dangerous-git.sh"

3. Register in Claude Code Settings

Read or create the settings file, then add the hook:

# Read existing settings or start fresh
if [ -f "$SETTINGS" ]; then
  CURRENT=$(cat "$SETTINGS")
else
  CURRENT='{}'
fi

# Use jq to safely add the PreToolUse hook (preserves existing config)
HOOK_PATH="$(pwd)/$HOOKS_DIR/block-dangerous-git.sh"
[[ "$ARGUMENTS" == "global" ]] && HOOK_PATH="$HOME/.claude/hooks/block-dangerous-git.sh"

echo "$CURRENT" | jq \
  --arg hook "$HOOK_PATH" \
  '.hooks.PreToolUse += [{"matcher": "Bash", "hooks": [{"type": "command", "command": $hook}]}]' \
  > "$SETTINGS"

echo "Hook registered in: $SETTINGS"

4. Verify

# Test the script directly
echo '{"command": "git push origin main"}' | bash "$HOOKS_DIR/block-dangerous-git.sh"
# Should exit 2 and print BLOCKED message to stderr

echo '{"command": "git status"}' | bash "$HOOKS_DIR/block-dangerous-git.sh"
# Should exit 0 (allowed)

5. Report

Git guardrails installed:

  Hook script:  .claude/hooks/block-dangerous-git.sh
  Settings:     .claude/settings.json
  Scope:        project

Blocked operations:
  git push / push --force
  git reset --hard
  git clean -f / -fd
  git branch -D
  git checkout . / git restore .

To customize: edit .claude/hooks/block-dangerous-git.sh
To test: echo '{"command": "git push"}' | bash .claude/hooks/block-dangerous-git.sh
To remove: delete the hook entry from .claude/settings.json

Blocked Operations

PatternWhy dangerous
git push / push --forceOverwrites remote history; hard to recover from
git reset --hardDiscards uncommitted work permanently
git clean -f / -fdDeletes untracked files/dirs permanently
git branch -DDeletes branch without merge check
git checkout . / git restore .Discards all working tree changes

Customization

Edit the dangerous_patterns array in the hook script to add or remove patterns. The script uses grep -E for matching, so regex patterns are supported.

Examples

/git-guardrails           ← project-level (.claude/settings.json)
/git-guardrails global    ← user-level (~/.claude/settings.json)

Edge Cases

  • jq not installed: Reports the required JSON change to make manually
  • Hook already exists: Detects duplicate matcher and skips re-adding
  • Global scope: Uses ~/.claude/ paths instead of project-local

Token Optimization

Expected range: 300–500 tokens

Early exit: Detects if hook already registered in settings before making changes.

Template-based: Script content is fixed — no LLM generation needed.

Patterns used: Early exit, template-based generation, Bash for system queries

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.