Skip to content
GitHub Copilot Train-the-Trainer
Trainer Guide

Session M7-01 trainer guide: Work effectively in the GitHub Copilot app

Duration: 1 hour
Audience: Product managers, product owners, business analysts, project managers, and delivery leads

Delivery objective

Show learners how to choose a session mode and attach approved source material in the GitHub Copilot app. Compare answers with and without those sources before creating the first Delivery Decision Studio brief.

One-hour plan

Time Segment
0:00-0:08 Orient learners to the app
0:08-0:18 Choose mode and context
0:18-0:28 Inspect capabilities and context usage
0:28-0:40 Compare unscoped and grounded answers
0:40-0:53 Create, revise, and check a working brief
0:53-1:00 Explain boundaries and launch the lab

Before the session

  • Check that every learner can open the GitHub Copilot app.
  • Check that the learner can start a chat and open the training repository as a project.
  • Check that the learner can see the current app navigation and available Customize items.
  • Open the starter initiative and reference brief.
  • Use only the fictional scenario during the demonstration.

Important

Stop if the app, repository, or required policy access is unavailable. A generic chatbot does not meet this session's goal.

Teach the app as a work environment

Show the current locations for:

  • Chats: exploratory work and interviews;
  • Projects: repository-grounded work;
  • My work: issues, pull requests, and active agent sessions;
  • Search: finding repository information;
  • Automations: recurring or on-demand tasks;
  • Customize: skills, agents, MCP servers, plugins, and canvases.

Labels can move. Teach what each area is for rather than a fixed click path.

Modes and context syntax

  • Interactive: use for interviews, source checks, and reviewed drafts.
  • Plan: use before a multi-step change so the learner can approve the plan.
  • Autopilot: reserve for work with agreed scope and permission to act without review of every action. It is not used for the Session M7-01 write.

GitHub documents @ for files, # for issues, and / for commands. Check the picker in the current app. Run /context before the first draft. Use /compact only when the learner needs to keep a long thread. Use /clear or /reset for the fresh-session test.

Context before prompting

Use a simple rule:

Use only permitted sources. Attach the smallest set that answers the question.

Demonstrate the difference between:

Explain this project.

and:

Use the repository overview and the open initiative notes. Explain the product,
the affected users, the current decision, and the missing information for a
product owner. Cite the source for each factual claim.

Ask learners which sources Copilot used and which statements remain uncertain.

Capability inspection

Open Customize and identify what is available in the learner's environment. Explain:

  • a skill provides a repeatable procedure;
  • an agent carries role-specific instructions and tools;
  • an MCP server lets Copilot read or act in another system;
  • a plugin is an installable package that can bundle skills, agents, hooks, MCP servers, or canvases;
  • a canvas provides visible shared state;
  • an automation repeats a task with a defined scope.

Do not build these customizations in this session. Show learners how to recognize them and explain that availability depends on plan, policy, and installation.

Demo: Create and revise a saved brief

  1. Start a fresh session, then ask for an initiative summary without attaching the source. Do not reuse the capability-check session that already read it.
  2. Mark unsupported statements.
  3. Attach lab/starter/initiative-overview.md and ask again with source references.
  4. Ask Copilot to complete initiative-context-brief-template.md.
  5. Review and commit the brief.
  6. Reveal lab/starter/late-evidence.md.
  7. Ask Copilot for a focused update and review the diff.
  8. Run /rubber-duck for an independent critique.
  9. Start a fresh session with only the saved brief and check that it can identify the outcome, facts, unresolved decisions, non-goals, and next owner action.

Stress the pattern:

Read -> propose -> review -> write -> verify

Point out that learners will use this same pattern in every later session.

Product notes verified October 5, 2026

Before delivery, check the current GitHub Copilot app navigation, plan requirements, policy controls, and Customize feature availability.

Official references: