linear-issue

name	linear-issue
description	Automate the complete Linear issue workflow - fetch issue, update status to In Progress, create branch, implement, test, create PR, and move to In Review

Linear Issue Workflow

Automates the complete workflow for working on Linear issues: fetch issue, update status, create branch, implement, test, create PR, and mark for review.

Usage

/linear-issue MEG-5

Or simply:

/linear-issue

(Will prompt you to select an issue)

Workflow

This skill automates the following steps:

1. Fetch Issue: Get issue details from Linear
2. Update to In Progress: Move issue to "In Progress" state
3. Create Branch: Create a new git branch based on issue identifier and title
4. Implement: Guide implementation according to issue requirements
5. Test: Verify the implementation works correctly
6. Create PR: Create a GitHub pull request with issue details
7. Update to In Review: Move issue to "In Review" state and link PR

Instructions

When this skill is invoked:

1. Parse the issue identifier from the arguments (e.g., "MEG-5"). If no argument provided, list recent backlog issues and ask user to select one.

2. Fetch the issue details from Linear using mcp__linear__Linear_GetIssue with the issue identifier.

3. Update issue status to "In Progress" using mcp__linear__Linear_TransitionIssueState with target state "In Progress".

4. Create a git branch:

   • Generate branch name from issue: feature/MEG-5-short-description (use kebab-case)
   • Run: git checkout -b <branch-name>
   • Confirm branch created successfully

5. Review the issue details with the user:

   • Display the issue title, description, and scope
   • Show the review checklist
   • Show files that will be changed
   • Ask user to confirm before proceeding with implementation

6. Implement the feature:

   • Follow the issue's objective and scope exactly
   • Use the TodoWrite tool to track implementation progress based on the issue's scope
   • Create/modify files as specified in the issue
   • Ensure code follows project conventions
   • Add proper error handling and type safety
   • Follow the review checklist items

7. Test the implementation:

   • Run relevant tests based on the issue's testing checklist
   • Verify no TypeScript errors: bun run type-check (if such script exists)
   • Test manually if needed
   • Ensure all success criteria are met
   • Fix any issues found during testing

8. Commit the changes:

   • Stage all changes: git add .
   • Create a descriptive commit message referencing the issue
   • Commit format: feat(scope): description\n\nFixes MEG-X\n\nCo-Authored-By: Claude Sonnet 4.5 <[e-mail]>

9. Push and create PR:

   • Push branch: git push -u origin <branch-name>
   • Create PR using gh pr create with:
     • Title: Issue title
     • Body: Include issue summary, implementation notes, testing done, and link to Linear issue
     • Use format:

       ## Summary
       [Brief description of changes]
       
       ## Implementation
       [Key implementation details]
       
       ## Testing
       - [x] Testing item 1
       - [x] Testing item 2
       
       ## Review Checklist
       [Copy from issue]
       
       Closes [Linear issue URL]
       
       🤖 Generated with [Claude Code](https://claude.com/claude-code)
       

10. Update issue to "In Review":

    • Use mcp__linear__Linear_TransitionIssueState with target state "In Review"
    • Add a comment to the Linear issue with the PR URL using mcp__linear__Linear_AddComment

11. Confirm completion:

    • Display summary of what was done
    • Show the PR URL
    • Show the updated Linear issue URL
    • Remind user that issue is now in review

Error Handling

• If issue not found, list available backlog issues
• If branch already exists, ask user if they want to switch to it or create a new one
• If tests fail, fix issues before creating PR
• If PR creation fails, provide manual instructions
• If Linear state transition fails, continue but warn user to update manually

Notes

• This skill assumes Linear states "In Progress" and "In Review" exist
• The skill will use fuzzy matching for state names if exact match fails
• Always verify implementation works before creating PR
• Never create PR with failing tests or TypeScript errors