What problem does it solve? Teams lack a consistent, objective bar for deciding whether a pull request is ready to merge, leading to subjective reviews, drive-by changes slipping in, and PRs merged without tests that fail on the base branch. ## Core Features & Use Cases - Fixed PR-level standard: Twelve scored properties covering single purpose, clean base, size under ~400 hand-written lines, justified hunks, tests that fail on base, dependencies, commits, and green CI on the PR head. - Structured scoring procedure: Fetches the full PR via gh pr view and gh pr diff, walks hunks rather than files, and emits a verdict line, fail list, score table, and unverified rows in a fixed output order. - Calibration workflow: Scores the last three merged PRs to reveal whether rows are stricter than the team wants when introducing the standard. - Use Case: Before opening a PR, ask whether it meets the bar; the skill checks each hunk against the title and tests, then reports exactly which rows fail with file:line evidence. ## Quick Start Ask the assistant to score the current pull request against the PR standard and report whether there is nothing to change.
