scope-drift

Compare stated intent to code changes and classify items as DONE, PARTIAL, NOT DONE, or CHANGED.

1|Updated Mar 19, 2026
One-click install
npx skills add https://github.com/lucasmccomb/ccgm --skill scope-drift
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: scope-drift
Source: https://github.com/lucasmccomb/ccgm/tree/main/modules/pr-review-toolkit/skills/scope-drift
Command: npx skills add https://github.com/lucasmccomb/ccgm --skill scope-drift

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Scope-drift identifies mismatches between the stated intent of work (PR body, commit messages, plan files, or task descriptions) and the actual changes in the diff so reviewers focus on the right problems and avoid merging unintended edits.

Core Features & Use Cases

  • Intent extraction: Collects intent from PR bodies, issue descriptions, commit messages, and plan/TODO files and normalizes them into actionable items.
  • Per-item classification: Compares each actionable item to the diff and marks it DONE, PARTIAL, NOT DONE, or CHANGED with file:line evidence.
  • Out-of-scope detection & gating: Flags files changed that are unrelated to intent, rates impact (HIGH/MEDIUM/LOW), and surfaces gating questions for HIGH-impact issues.
  • Use cases: Run at the start of a PR review to verify deliverables, before claiming tasks as complete, or when a diff seems larger than expected.

Quick Start

Run a scope-drift audit comparing the PR body, commits, and any plan files to the diff and produce a Scope Drift Audit report with item classifications and file:line evidence.

Frequently Asked Questions about scope-drift

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I detect scope drift in a pull request before reviewing the code?

To detect scope drift in a pull request, compare the stated intent from the PR body and commit messages against the actual diff. This process classifies deliverables as DONE or PARTIAL and flags out-of-scope files with impact ratings.

What is scope drift in code review and how does intent analysis find it?

Scope drift in code review occurs when actual code changes mismatch the stated PR intent. Intent analysis finds it by extracting deliverables from plan files or commit messages and verifying each item against the diff with file:line evidence.

How do I audit a git diff for out-of-scope changes and unintended edits?

You can audit a git diff for out-of-scope changes by comparing the code changes to the original task description. The diff audit flags unrelated files, rates their impact HIGH or LOW, and generates gating questions for HIGH-impact issues.

Can I verify task completion in a multi-file PR using commit messages and plan files?

Yes, you can verify task completion in a multi-file PR by normalizing intent from commit messages and plan files into actionable items. The review then checks each item against the diff to produce DONE, PARTIAL, or NOT DONE classifications.

What is the best way to generate gating questions for high-impact PR review issues?

The best way to generate gating questions for high-impact PR review issues is to flag out-of-scope changed files with impact ratings. When HIGH-impact issues are detected during the diff audit, targeted gating questions are surfaced automatically.