Copilot Suggestions

Suggestions accelerate a first draft when you have already established the pattern. They do not define product requirements or architecture for you.

#Useful starting points

A function name/header, a few comments, repetitive implementation, a test skeleton or a data mapping gives the model a concrete direction.

#Types of suggestion

Inline completion extends a statement, object, chained call or local implementation. Multi-line completion can produce a larger CRUD, test or mapping skeleton when context is clear. If the first candidate is unsuitable, inspect alternatives rather than accepting it automatically.

#Provide structure

tsCopy
async function fetchUserProfile(userId: string) {
  // 1. call API
  // 2. handle error
  // 3. return normalized object
}

tsCopy
// load user settings from local storage
// return default settings if missing
export function loadUserSettings() {
  // Copilot continues here
}

These are prompting skeletons, not complete implementations. Clear types, parameters and variable names reduce guessing.

#Review the draft

Boilerplate, test skeletons, well-defined helpers and type/data mappings are suitable uses. Permissions, critical branches and security decisions still need human review. Treat completion as a first-draft accelerator rather than a final implementation generator.

#Related pages

• Copilot Chat
• Setup guide
• Copilot overview