guide-mode

Require user approval before state-changing operations during code walkthroughs.

3|1|Updated Mar 4, 2026
One-click install
npx skills add https://github.com/omaclaren/agent-skills-public --skill guide-mode
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: guide-mode
Source: https://github.com/omaclaren/agent-skills-public/tree/main/skills/guide-mode
Command: npx skills add https://github.com/omaclaren/agent-skills-public --skill guide-mode

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Guide mode prevents autonomous edits and unexpected mutations by prioritizing step-by-step walkthroughs, explanations, and user approval so learners and reviewers retain ownership and understanding of code and results.

Core Features & Use Cases

  • Interactive Walkthroughs: Drive sessions where the agent explains code, runs safe read-only checks, and proposes incremental changes for user approval.
  • User Sign-off Before Mutations: Require explicit permission before writing files, committing, pushing, or performing long-running operations.
  • Use Case: A developer learning a new codebase asks to inspect a function; the agent explains the function line-by-line, suggests a minimal change, and waits for approval before applying it.

Quick Start

Invoke /guide-mode and ask for a step-by-step walkthrough of the file or task you want to inspect.

Frequently Asked Questions about guide-mode

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

FAQPage Schema
How do I prevent an AI agent from making autonomous edits during a code review?

Interactive walkthroughs prevent autonomous edits by requiring user sign-off before applying any state-changing operations. Guide mode blocks file writes, commits, and pushes until explicit permission is granted, ensuring reviewers retain ownership of the codebase.

What is the best way to walk through a new codebase line-by-line with an AI?

Step-by-step walkthroughs enable an agent to explain code line-by-line, run safe read-only checks, and propose incremental changes. This interactive exploration approach helps developers learning a new codebase understand functions deeply before approving any modifications.

How do I require explicit approval before an AI commits or pushes code changes?

User sign-off before mutations requires explicit permission before writing files, committing, pushing, or performing long-running operations. By invoking guide mode, state-changing operations are blocked until the user provides direct approval for each proposed change.

Can I use interactive walkthroughs for debugging sessions without losing control of my code?

Yes, interactive walkthroughs for debugging sessions prevent unexpected mutations by prioritizing explanations and incremental approval. The agent explains the issue, suggests minimal changes, and waits for explicit user sign-off before applying any modifications to the code.

Does guide mode work for learning sessions where I need to inspect a function safely?

Guide mode works for learning sessions by enabling safe, read-only inspections where the agent explains a function and proposes changes. It prevents autonomous mutations, allowing learners to inspect code interactively and approve modifications incrementally.