batch-orchestration

What problem does it solve? Applying the same change pattern across dozens of files (migrations, codemods, refactors) is slow and error-prone when done sequentially, and risky when done in one giant edit. This Skill splits the work into independent units and runs them in parallel with isolated worktrees and per-unit verification. ## Core Features & Use Cases - Research and Decomposition: Scans the repo to find every matching instance, then splits the work into 5-30 independent, self-contained, verifiable units. - Plan Approval Gate: Presents the full unit list and waits for explicit confirmation before spawning any agents. - Parallel Execution with Isolation: Each unit runs in its own git worktree with a dedicated background agent that implements the change, runs tests, and opens a PR. - Use Case: Convert 18 React class components to function components — the Skill finds all 18, gets your approval, then runs 18 parallel agents that each convert one file, run its tests, and open a PR. ## Quick Start Ask the AI to run /batch followed by a change pattern, such as converting all React class components to function components across the codebase.