check-drift

Detects implementation drift by comparing built code to locked specs and generating a correction manifest.

3|Updated Jan 25, 2026
One-click install
npx skills add https://github.com/kapilvirenahuja/garura --skill check-drift-kapilvirenahuja
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: check-drift
Source: https://github.com/kapilvirenahuja/garura/tree/main/core/components/skills/check-drift
Command: npx skills add https://github.com/kapilvirenahuja/garura --skill check-drift-kapilvirenahuja

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill identifies where implemented changes no longer match locked specs, so teams can correct alignment before delivery and avoid hidden rework caused by silent divergence.

Core Features & Use Cases

  • Drift classification with RCA: Detects context, spec, technology, design, and evidence drift, explaining why each divergence happened.
  • Actionable outputs for correction: Generates drift reports, ADR drafts (for deliberate decisions), debt catalogs, memory items for LTM promotion, and prioritized spec/code recommendations.
  • Human-reviewed write-back: Produces a spec-correction-manifest.yaml that proposes precise spec updates without auto-modifying locked artifacts.

Quick Start

Ask the assistant to run a drift check for a specific scope by giving an issue number or epic ID so it can generate a scoped drift report and spec-correction manifest.

Frequently Asked Questions about check-drift

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

FAQPage Schema
How do I detect spec drift between my implementation and locked design documents?

Spec drift is detected by mapping git changes and evidence to locked design and product specifications, classifying five specific divergence types with root cause analysis to identify where implementation no longer matches the original requirements.

What is the best way to validate technical implementation against product specifications before delivery?

Validating implementation against product specifications requires scoped drift checks for branch-based or issue-based units, generating a spec-correction-manifest that proposes precise spec updates without auto-modifying locked artifacts for human review.

How do I generate an architecture decision record for deliberate technical deviations during development?

Architecture decision records are drafted as actionable outputs during drift analysis, specifically capturing deliberate decisions where the implementation diverged from the locked specs, alongside generating debt catalogs and memory items.

Can I run an ad-hoc drift analysis on a specific epic ID without modifying locked artifacts?

Yes, ad-hoc drift analysis can be scoped to an epic ID by mapping git changes to spec artifacts, producing a spec-correction-manifest.yaml that proposes updates for human review without automatically modifying the locked artifacts.

What types of implementation drift can be classified during a post-implementation review?

Implementation drift is classified into five types during post-implementation reviews: context drift, spec drift, technology drift, design drift, and evidence drift, each with root cause analysis explaining why the divergence occurred.