/fix-and-create-pr-linear-issue

Fix and Create PR for Linear Issue

Input: $ARGUMENTS

Step 1: Find the Issue

Determine if the input is an issue ID or search text:

• Issue ID pattern (matches [A-Z]+-\d+, e.g. EO-42): Use mcp__claude_ai_Linear__get_issue to fetch the issue directly by its identifier.

• Search text (anything else): First, ask the user which team and project to search within:

  1. Fetch teams and projects in parallel using mcp__claude_ai_Linear__list_teams and mcp__claude_ai_Linear__list_projects.
  2. Use a single AskUserQuestion call to ask:
     • Team: "Which team?" (list all teams). If only one team exists, auto-select it and skip this question.
     • Project: "Which project?" (list all projects). If only one project exists, auto-select it and skip this question.
  3. Search for issues matching the text, scoped to the selected team/project, with limit: 10. If multiple results, present them to the user via AskUserQuestion and ask which one to fix. If no results, tell the user and stop.

  Important: AskUserQuestion requires at least 2 options per question. If a question would have only 1 option, auto-select that option and omit it from the call.

Step 2: Update Linear Status to "In Progress"

Immediately use mcp__claude_ai_Linear__save_issue to move the issue status to "In Progress". This must happen before any confirmation prompt or agent work, so that subsequent /loop iterations do not pick up the same issue.

If the issue is already "In Progress", this is a no-op — proceed without error.

If the status update fails for another reason (e.g., invalid state transition), warn the user but continue.

Step 3: Confirm with User

Display the issue details:

• ID: [issue identifier]
• Title: [issue title]
• Description: [issue description, truncated if very long]

Ask the user: "Fix this issue and create a PR?" with options Yes / No.

If No, use mcp__claude_ai_Linear__save_issue to revert the issue status back to its previous status, then stop.

Step 4: Spawn the Issue PR Creator Agent

Use the Task tool with isolation: "worktree" and subagent_type: "linear-integration:issue-pr-creator" to spawn the agent.

Pass the full issue context in the prompt:

Fix this Linear issue and create a PR:

- **Issue ID:** [id]
- **Title:** [title]
- **Description:** [full description]
- **Labels:** [labels if any]
- **Priority:** [priority if any]

Work autonomously: explore the codebase, implement the fix, create tests, run verification, run /simplify, re-verify,
post a detailed summary to the Linear issue comments, commit, push, create a GitHub PR, and post the PR link to the
Linear issue.

Step 5: After Agent Completes

On Success

1. Report the results to the user (PR URL, summary of changes).
2. Use mcp__claude_ai_Linear__save_issue to move the issue status to "Done".
3. Clean up the worktree: run git worktree remove <worktree-path> using the path returned by the agent. If cleanup fails, warn the user but do not treat it as an error.

On Failure

1. Report the failure to the user.
2. Use mcp__claude_ai_Linear__save_comment to post an error comment on the issue:

   ## Automated Fix Failed
   
   The autonomous fix attempt encountered an error:
   
   [error details]
   
   The branch may contain partial work. Manual intervention is needed.
   

3. Use mcp__claude_ai_Linear__save_issue to move the issue status back to "Todo".
4. Clean up the worktree: run git worktree remove <worktree-path> --force. If cleanup fails, warn the user.