breakup-pr
Break Up PR into Vertical Feature Slices
You are splitting a large PR into vertical feature slices — each slice is an independently mergeable, CI-green stacked PR delivered via Graphite (see the graphite skill for gt commands).
All slices are gated behind a single shared feature flag so the feature ships atomically when the flag is enabled.
Why Vertical Slices?
Large PRs are hard to review, hard to reason about, and risky to merge. Vertical slicing solves this by breaking work into small, focused changes that are easier to understand and safer to ship.
Easier code review:
- Reviewers can focus on one concern at a time (e.g., "just the migration" or "just the API")
- Each PR tells a clear story — what changed and why is obvious from the diff
- Smaller diffs mean faster reviews and higher-quality feedback
- Comments and discussions stay focused instead of sprawling across unrelated changes
More from lightdash/lightdash
developing-in-lightdash
Use when working with Lightdash YAML files, dbt models with Lightdash metadata, the lightdash CLI (deploy, upload, download, preview, lint, sql, set-warehouse), or creating/editing charts, dashboards, metrics, and dimensions as code
91frontend-style-guide
Apply the Lightdash frontend style guide when working on React components, migrating Mantine v6 to v8, or styling frontend code. Use when editing TSX files, fixing styling issues, or when user mentions Mantine, styling, or CSS modules.
64ld-permissions
Guide for Lightdash's CASL-based authorization system. Use when working with scopes, custom roles, abilities, permissions, ForbiddenError, authorization, or access control. Helps with adding new scopes, debugging permission issues, understanding the permission flow, and creating custom roles.
48debug-local
Debug the Lightdash app using PM2 logs, Spotlight traces, and browser automation. Use when investigating issues, tracking down bugs, understanding request flow, or correlating frontend actions with backend behavior.
23har-replay
Replay a HAR file as a mock backend to reproduce frontend performance issues with production data. Use when asked to replay a HAR file, reproduce a dashboard with a HAR, or test frontend performance with captured traffic.
23investigate-pylon
Investigate a Pylon support ticket. Use when asked to investigate, triage, or look into a Pylon issue. Classifies the ticket, searches for existing GitHub/Linear issues, searches docs, inspects code if needed, and drafts a customer reply.
21