Mentor Mode
Coaches instead of coding: it looks through the codebase to understand context, then challenges your assumptions with questions and the 5 Whys rather than editing files or handing you a solution.
Overview
Mentor Mode is built for a specific moment: when you're mid-feature and want a second opinion, not a pair programmer who takes over the keyboard. It never makes code edits, only asks questions and points out what it thinks you're missing.
How it works
- It reads the codebase, searches for related files and function usages to understand your actual context before saying anything about your approach.
- It asks questions to clarify your understanding of the problem and the solution you've proposed, looking specifically for unexamined assumptions.
- It uses Socratic questioning and the 5 Whys to push past the surface explanation and get to the underlying reasoning, rather than accepting the first answer.
- When it spots an unsafe practice or a real problem, it says so plainly and explains the long-term cost of the shortcut, instead of hedging or apologizing for pointing it out.
- It stays concise on purpose. The instructions explicitly favor being clear over being thorough, since the goal is to prompt your own thinking, not to write a report.
Examples
"I'm going to skip writing tests for this migration script since it's a
one-time run."
The agent asks what happens if the script fails halfway through production data, and whether "one-time" has held true for scripts like this before, before you commit to skipping tests.
Installation
You need VS Code with the GitHub Copilot extension.
- Create the agents folder and download the file into your repo:
mkdir -p .github/agents curl -o .github/agents/mentor.agent.md \ https://raw.githubusercontent.com/github/awesome-copilot/main/agents/mentor.agent.md - Reload VS Code, or open a new Copilot Chat; custom agents under
.github/agents/are picked up automatically. - Pick this agent from the mode picker in Copilot Chat and describe your approach to confirm it questions you instead of editing files.
Related assets
Accessibility Expert
Reviews code and designs against WCAG 2.1/2.2, covering semantics, keyboard and focus behavior, forms, media and dynamic SPA updates, and ships framework-specific examples plus a CI setup for automated checks.
Address Comments Agent
Works through pull request review comments one at a time: applies the minimal fix for each, pushes back on comments that don't make sense, adds test coverage, and commits with a descriptive message before moving to the next comment.
ADR Generator
Turns a decision discussion into a numbered Architecture Decision Record: gathers the missing context, assigns the next sequential ADR number, and writes a structured markdown file with consequences, alternatives and rejection reasons.