Does this skill require manual code annotations?

No, while it supports optional annotations like '// Implements: PRD-1', it primarily uses heuristics such as test file analysis and function naming to detect alignment.

Can it identify extra work that wasn't requested?

Yes, it specifically flags 'Extra' or 'Scope Creep' items when it detects new functionality that does not map back to any defined requirement in your PRD or SDD.

How does it log drift decisions?

The skill automatically appends findings and user decisions to a dedicated 'Drift Log' section in your project's specification README for permanent traceability.

What is drift detection in software development?

Drift detection is the process of identifying when the actual code implementation deviates from the original design documents, requirements, or project plans.

Drift Detection

Name: Drift Detection
Author: rsmdt

byrsmdt

•

158

•

生産性とワークフロー

Monitors and identifies discrepancies between project specifications and code implementation to ensure architectural alignment.

The Drift Detection skill acts as a specification alignment specialist for Claude Code, continuously monitoring implementation phases to identify scope creep, missing features, and contradictions. By analyzing PRDs, design documents, and project plans against actual code and tests, it provides developers with actionable insights and logs alignment decisions directly to project documentation for full traceability. This ensures that development remains focused on defined goals while providing a conscious process for acknowledging and documenting necessary deviations.

主な機能

01158 GitHub stars

02Automated comparison between requirements and current implementation

03Heuristic-based detection using test analysis and code patterns

04Integrated drift logging to README files for audit trails

05Categorization of drift into Scope Creep, Missing, Contradicts, and Extra

06Interactive decision prompts for resolving alignment issues

ユースケース

01Verifying phase completion against acceptance criteria during a sprint

02Auditing code changes during PR reviews to ensure alignment with design specs

03Preventing unplanned features from bloating the codebase without documentation

主な機能

01158 GitHub stars

02Automated comparison between requirements and current implementation

03Heuristic-based detection using test analysis and code patterns

04Integrated drift logging to README files for audit trails

05Categorization of drift into Scope Creep, Missing, Contradicts, and Extra

06Interactive decision prompts for resolving alignment issues

ユースケース

01Verifying phase completion against acceptance criteria during a sprint

02Auditing code changes during PR reviews to ensure alignment with design specs

03Preventing unplanned features from bloating the codebase without documentation