Skip to main content
Agent skill

Safety Guard

Guard against destructive operations with three modes: Careful intercepts dangerous commands (rm -rf, git push --force, DROP TABLE) for confirmation, Freeze locks writes to one directory, and Guard combines both. Advisory: the agent applies these rules itself; no hook ships with it. Use when working on production systems, running agents autonomously, restricting edits to a directory, or during migrations, deploys, and data changes.

Get started

Get the complete skill folder, including its scripts and reference files.

·

When to use this skill

Use Safety Guard when an AI agent needs a reusable SKILL.md workflow for this job: Guard against destructive operations with three modes: Careful intercepts dangerous commands (rm -rf, git push --force, DROP TABLE) for confirmation, Freeze locks writes to one directory, and Guard combines both. Advisory: the agent applies these rules itself; no hook ships with it. Use when working on production systems, running agents autonomously, restricting edits to a directory, or during migrations, deploys, and data changes.

When not to use it

Skip Safety Guard when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

Where will you use it?

Install scope
Claude Code folder setup
mkdir -p ~/.claude/skills/safety-guard
# Save SKILL.md to ~/.claude/skills/safety-guard/SKILL.md
  1. Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/safety-guard/SKILL.md. For a shared project, commit the folder at .claude/skills/safety-guard/ instead.
  2. Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
  3. For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.

Full install guide for Claude, Cursor, and Codex →

Try a starter prompt

After installing the skill, paste this starter prompt into your agent to verify it triggers on a small task.

Starter prompt
Use the Safety Guard skill for this task: When working on production systems. Confirm any required inputs or files before making changes.

What supports this recommendation?

74/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.

Documentation
82/100
Practical detail
74/100
Source evidence
63/100
Source signals
72/100
Model compatibility evidence
  • claudeinferred

    The skill text mentions Claude or a closely associated term. This is keyword-derived compatibility; no hands-on model test is recorded.

  • chatgptuntested

    No model-specific signal or recorded compatibility test was found.

  • geminiuntested

    No model-specific signal or recorded compatibility test was found.

  • copilotuntested

    No model-specific signal or recorded compatibility test was found.

  • llamauntested

    No model-specific signal or recorded compatibility test was found.

  • perplexityuntested

    No model-specific signal or recorded compatibility test was found.

  • mistraluntested

    No model-specific signal or recorded compatibility test was found.

  • grokuntested

    No model-specific signal or recorded compatibility test was found.

Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.

What this skill does

# Safety Guard — Prevent Destructive Operations

## When to Use

- When working on production systems - When agents are running autonomously (full-auto mode) - When you want to restrict edits to a specific directory - During sensitive operations (migrations, deploys, data changes)

## How It Works

Three modes of protection. They are instructions the agent follows, not enforced blocks (see [Enforcement](#enforcement)).

### Mode 1: Careful Mode

Before it runs a destructive command, the agent stops and warns:

``` Watched patterns: - rm -rf (especially /, ~, or project root) - git push --force - git reset --hard - git checkout . (discard all changes) - DROP TABLE / DROP DATABASE - docker system prune - kubectl delete - chmod 777 - sudo rm - npm publish (accidental publishes) - Any command with --no-verify ```

When detected: shows what the command does, asks for confirmation, suggests safer alternative.

### Mode 2: Freeze Mode

Locks file edits to a specific directory tree, for example: "freeze edits to `src/components/`".

The agent does not Write/Edit outside `src/components/`, and explains why when a change would need it. Useful when you want an agent to focus on one area without touching unrelated code.

### Mode 3: Guard Mode (Careful + Freeze combined)

Both protections active, for example: "guard mode, write only to `src/api/`, read anything".

The agent can read anything but only writes to `src/api/`, and asks before any destructive command.

### Unlock

Tell the agent to stop applying the rules.

## Enforcement

This skill is advisory. ECC ships no `/safety-guard` command and no hook for it, so nothing blocks a tool call that breaks these rules. For blocking that does not depend on the agent, use Claude Code [permission deny rules](https://code.claude.com/docs/en/permissions) (for example `Bash(rm -rf *)`) or a [PreToolUse hook](https://code.claude.com/docs/en/hooks).

## Integration

- Enable by default for `codex -a never` sessions - Pair with observability risk scoring in ECC 2.0

Intended uses

  • When working on production systems
  • When agents are running autonomously (full-auto mode)
  • When you want to restrict edits to a specific directory
  • During sensitive operations (migrations, deploys, data changes)

Implementation notes

  • Review the source instructions and adapt inputs before running the workflow.

Try it, then tell us how it went

A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.