linear-issue-driven-development

Automating Linear issue to GitHub pull request workflows involves using a headless agent to clone the repository, implement changes, push branches, open PRs, monitor CI, and resolve review threads until mergeable. This Skill orchestrates that entire process end to end.

Issue-driven development automation for headless agents is a process where an agent fetches a ready Linear issue, locks it to prevent duplicate runs, implements the smallest acceptable change, and leaves a clean, reviewable PR with correct final status labels.

Yes, you need separate authentication tokens. The workflow requires Linear API authentication to fetch and update issue statuses, and a GitHub token to clone repositories, push commits, open pull requests, and resolve review threads.

Automated CI monitoring watches GitHub pull request checks until they turn green. Once CI passes, the workflow proceeds to resolve any review threads, ensuring the pull request reaches a fully mergeable state without manual intervention.

To prevent duplicate concurrent runs on the same Linear issue, the workflow uses a lock marker and label transitions. This ensures only one headless agent processes a specific issue at a time, managing failure states through dedicated label updates.

The best way to manage failure states in automated pull request workflows is through failure-state label management. This Skill applies specific label transitions on Linear issues when errors occur, ensuring clear tracking and preventing duplicate processing attempts.