gateguard

Block code edits until investigation verifies imports, schemas, and commands.

Updated Jun 24, 2026

Search Tags:#investigation #schema validation #code safety #gatekeeping #pre-edit #destructive commands #codex hooks

Ready to run

npx skills add https://github.com/mostafa-ismail-2004/codex-plugin --skill gateguard-mostafa-ismail-2004

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: gateguard
Source: https://github.com/mostafa-ismail-2004/codex-plugin/tree/main/skills/gateguard
Command: npx skills add https://github.com/mostafa-ismail-2004/codex-plugin --skill gateguard-mostafa-ismail-2004

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

GateGuard prevents AI agents from guessing before making changes by forcing a fact-gathering step first. It is designed to reduce risky edits, bad assumptions, and destructive commands in codebases where accuracy depends on understanding imports, schemas, and user intent.

Core Features & Use Cases

Pre-edit investigation gate: Blocks the first Edit, MultiEdit, or Write attempt until concrete facts are gathered.
Destructive command protection: Adds stricter checks for risky Bash actions such as deletes, force resets, and force pushes.
Workflow validation: Helps teams enforce source tracing, schema inspection, and verbatim instruction quoting before code changes.
Use case: Ideal when modifying shared modules, data-processing code, or any workflow where one change can affect multiple files.

Quick Start

Use the gateguard skill to enforce a fact-gathering gate before any edit or destructive command in this project.

Frequently Asked Questions about gateguard

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I prevent an AI agent from editing code before verifying facts?▼

To prevent premature code editing, you can enforce a pre-edit investigation gate that blocks Edit, MultiEdit, or Write attempts until traceable file impact analysis and schema confirmation are gathered.

Can I block destructive shell commands like force pushes and deletes during automated code changes?▼

Yes, you can block destructive shell commands by applying stricter checks that intercept risky Bash actions such as deletes, force resets, and force pushes before they execute.

What is pre-edit investigation for code safety and how does it work?▼

Pre-edit investigation for code safety works by requiring concrete fact-gathering, such as verifying imports, shared modules, and data schemas, before allowing any modifications to proceed.

How do I enforce schema validation and user-instruction quoting before modifying shared modules?▼

You can enforce schema validation and verbatim user-instruction quoting by implementing a gatekeeping workflow that requires these confirmations before permitting edits to shared modules.

Does GateGuard work with codebases that have complex imports and data schemas?▼

Yes, GateGuard is designed for codebases with complex imports and data schemas, requiring traceable file impact analysis and schema confirmation to safely permit edits.

What are the limitations of using a pre-edit investigation gate for AI code modifications?▼

A limitation of using a pre-edit investigation gate is that it blocks the first edit attempt until fact-gathering is complete, which adds a mandatory step before any code changes can proceed.