GitHub Copilot lessons: workflows, costs, mistakes, prompts
Reference lessons to use after the step-by-step tutorial.
Workflows
Issue to PR
- Write a clear issue with acceptance criteria.
- Assign it to Copilot.
- Review the PR and CI.
Credits explained
Base credits plus a flex allotment each month. After that, extra usage is billed against your budget at $0.01 per credit.
Prompt library
Explain what this function does, list its edge cases, and point out one bug risk. Don't rewrite it.Use-case playbooks
Issue to PR
- Write acceptance criteria in the issue.
- Assign it to Copilot.
- Review the PR and CI output.
Test backfill
- Name the function and cases.
- Forbid changes to the function.
- Run the tests locally.
Code review pass
- Ask for risks, not style nits.
- Check each flagged line yourself.
Animated walkthrough
Animated walkthrough
Press play, or pick a chapter. Each chapter animates its steps.
Troubleshooting
I hit a limit. What now?
Agent mode, code review, Copilot in the cloud, the Copilot CLI and the Copilot app use AI credits; 1 credit = $0.01 (GitHub). Start a new chat with a short summary instead of dragging a long one.
The answer cites a page that doesn't say that.
Treat the claim as unverified. Ask: quote the exact sentence from that page that supports this, or say it isn't there. Then open the page yourself.
It forgot my earlier instructions.
Long chats push early text out of context. Put standing instructions where GitHub Copilot keeps them (Repository-wide custom instructions live in .github/copilot-instructions.md; path-specific instructions apply to matching files (GitHub docs).) and start a fresh chat.
It agrees with everything I say.
Ask it to argue the other side and list the strongest evidence against your view, with sources.
How do I keep costs down?
Pro $10/mo, Pro+ $39/mo, Max $100/mo, each with base plus flex credits. Set an extra-usage budget; alerts come at 75%, 90% and 100%. A $0 budget means no overage (GitHub).
Cost control
Agent mode, code review, Copilot in the cloud, the Copilot CLI and the Copilot app use AI credits; 1 credit = $0.01 (GitHub).
- Test with the free tier on real tasks for a week first.
- Set a calendar reminder 2 days before any renewal.
- Keep one paid assistant, not three, unless each earns its place.
Prompt library: bad vs good
Bad: help me
Good: Write pytest tests for parse_invoice() in invoices.py: the happy path, a missing total, a negative amount, and a DD/MM date. Don't change the function.
Why: situation, task, format and limits.
Bad: research this
Good: Find the 5 most relevant official sources from the last 12 months on [topic]. For each: title, date, link, and one sentence it actually says.
Why: asks for dated, quotable, checkable sources.
Bad: make it better
Good: Rewrite this for a busy customer: max 100 words, one ask, no jargon. Then list what you cut.
Why: a reader, a length and an audit of changes.
Bad: should I do X?
Good: List the 3 strongest reasons for and against X for someone in [situation]. Mark which reasons depend on facts I should check.
Why: balanced, and separates facts from judgement.
Privacy and safety
Install the GitHub Copilot extension (VS Code, JetBrains or Visual Studio) and sign in with the same GitHub account.
- Don't paste passwords, ID numbers, card numbers or other people's private details.
- Assume the company can read and keep your chats unless its official policy says otherwise.
- For health, money or legal decisions, confirm with a professional or an official source.
Failure modes to watch
- Tests that assert the current (buggy) behaviour.
- Silent edits to unrelated files.
- Secrets pasted into chat.
- Credit burn from long agent loops.
When to use it vs other bots
| Use GitHub Copilot for | Consider another bot when |
|---|---|
|
See: ChatGPT · Claude · Gemini · Perplexity · Copilot · comparison |
Leveled tutorials
Where these facts come from
- github.com/features/copilot/plans: plans, AI credits, flex allotments, budgets, alerts
Checked Oct 29, 2026. Plans change, so confirm on the official page before you pay.
Next: Cursor tutorial →