- Home
- Skills
- Agents & Orchestration
- Destructive Git Guard
Works with the AI Agents you already use
Destructive Git Guard
Prevents AI agents from executing destructive Git or filesystem commands without explicit confirmation and impact analysis.
$5
Destructive Git Guard
Example session with this skill installed
Clean up my workspace so I can start fresh on the login feature.
- Read your context and instructions
- Compiled the destructive git guard
I've checked the status. Running git clean -fd will permanently delete 4 untracked files including config/secrets.tmp. You have 3 uncommitted changes in auth.ts. Shall I stash these and then run the clean command, or do you explicitly authorize deleting these files permanently?
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
AI agents with terminal access can execute destructive commands like git reset --hard or rm -rf, permanently deleting hours of uncommitted work without warning. Standard agents often prioritize efficiency over safety, leading to irreversible data loss in environments without remote backups.
What it does
- Intercepts high-risk commands including
git reset,checkout,clean,push --force, andrm. - Forces a mandatory
git statuscheck to audit the working tree before any destructive proposal. - Calculates and reports the exact scope of potential data loss to the human operator.
- Blocks execution until explicit, per-command confirmation is received for the specific action.
- Recommends non-destructive alternatives like
git stashor branch creation before proceeding with deletions.
Why this beats prompting it yourself
Simple system prompts for "safety" are often ignored during complex task execution. This skill enforces a specific pre-flight checklist and state-check workflow that makes silent data loss architecturally difficult for the agent.
Use cases
- Protecting local-only projects that lack a remote origin or recent commits.
- Managing agents running in external terminals or CLI wrappers outside of IDE undo-history.
- Preventing accidental history overwrites during complex rebase or force-push operations.
- Ensuring untracked files are not nuked during automated workspace cleanup.
Known limitations
Requires the agent to have access to standard Git and shell binaries to perform state checks. It cannot prevent data loss if the agent is explicitly instructed by the user to bypass all safety protocols.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires Git and standard shell environment. Safety checks rely on the agent's ability to run git status and git diff --stat before proceeding.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 3 months ago
- Software
Git