What problem does it solve? Manually creating pull requests in a monorepo involves analyzing changes across packages, writing conventional commit messages, generating branch names, applying the right labels, and formatting descriptions — a repetitive, error-prone process. This Skill automates the entire PR creation workflow for the ai_clients_dart monorepo. ## Core Features & Use Cases - Automated change analysis: Inspects uncommitted changes and existing commits, categorizes them by package, and plans conventional commits grouped by type and scope. - Structured PR generation: Produces conventional commit titles under 70 characters and descriptions with Summary, Details, Breaking Changes, References, and Test Plan sections that feed changelog generation. - Label and assignee mapping: Maps changed package directories to p: labels, commit types to t: labels, and resolves assignees from CODEOWNERS. - Use Case: After implementing a new streaming feature in openai_dart and fixing a bug in chromadb, ask the assistant to create a PR — it commits the changes with proper conventional messages, pushes a branch, and opens a labeled PR via gh. ## Quick Start Ask the assistant to create a pull request for your current changes, optionally passing --draft or --base to target a different branch.
