
using-unwind
by cliftonc
Claude skills to deeply analyse and document a legacy codebase to rebuild with AI
SKILL.md
name: using-unwind description: Use when starting any reverse engineering task - establishes how to find and use Unwind skills for codebase analysis, service mapping, and documentation allowed-tools:
- Read
- Grep
- Glob
Using Unwind
Overview
Unwind provides structured skills for reverse engineering codebases. Produces complete, machine-readable documentation with source links.
Principles
See analysis-principles.md:
- Completeness: Document ALL items (30 tables = 30 documented)
- Machine-readable: Use actual code, SQL, mermaid - not markdown recreation
- Link to source: GitHub links with line numbers where possible
- No commentary: Facts only, no speculation or recommendations
Workflow
start → architecture.md (+ repo info for linking)
│
unwinding-codebase → dispatches layer specialists
│
├── analyzing-database-layer → database/
├── analyzing-domain-model → domain-model/
├── analyzing-service-layer → service-layer/
├── analyzing-api-layer → api/
├── analyzing-messaging-layer → messaging/ (if present)
├── analyzing-frontend-layer → frontend/ (if present)
├── analyzing-unit-tests → unit-tests/
├── analyzing-integration-tests → integration-tests/
└── analyzing-e2e-tests → e2e-tests/
│
verifying-layer-documentation → gap detection (parallel per layer)
│
└── Outputs gaps.md per layer (missing items only)
│
completing-layer-documentation → gap completion (parallel per layer)
│
├── Reads gaps.md work list
├── Adds missing documentation
└── Deletes gaps.md when done
│
synthesizing-findings → REBUILD-PLAN.md (strategic rebuild approach)
Skills
Core Flow
| Skill | Output |
|---|---|
start | architecture.md |
unwinding-codebase | Orchestrates layer analysis |
verifying-layer-documentation | gaps.md per layer (work list) |
completing-layer-documentation | Fills gaps, deletes gaps.md |
synthesizing-findings | REBUILD-PLAN.md |
Layer Specialists
| Skill | Output |
|---|---|
analyzing-database-layer | database.md |
analyzing-domain-model | domain-model.md |
analyzing-service-layer | service-layer.md |
analyzing-api-layer | api.md |
analyzing-messaging-layer | messaging.md |
analyzing-frontend-layer | frontend.md |
Testing Specialists
| Skill | Output |
|---|---|
analyzing-unit-tests | unit-tests.md |
analyzing-integration-tests | integration-tests.md |
analyzing-e2e-tests | e2e-tests.md |
Output Structure
docs/unwind/
├── architecture.md
├── layers/
│ ├── database/
│ │ ├── index.md
│ │ ├── schema.md
│ │ ├── repositories.md
│ │ └── verification.md
│ ├── domain-model/
│ │ ├── index.md
│ │ ├── entities.md
│ │ └── verification.md
│ ├── service-layer/
│ │ ├── index.md
│ │ ├── services.md
│ │ ├── formulas.md
│ │ └── verification.md
│ ├── api/
│ │ ├── index.md
│ │ ├── endpoints.md
│ │ └── verification.md
│ └── [other layers...]
└── REBUILD-PLAN.md
Each layer is a folder with index.md + section files for incremental writes.
Quick Start
Use unwind:start- Review
docs/unwind/architecture.md Use unwind:unwinding-codebaseUse unwind:verifying-layer-documentation(runs parallel verification)Use unwind:synthesizing-findings
Note: Step 4 (verification) is integrated into unwinding-codebase but can also be run independently to re-verify existing documentation.
Refresh Mode
Re-run any skill to update documentation. Changes highlighted in ## Changes Since Last Review section.
Score
Total Score
Based on repository quality metrics
SKILL.mdファイルが含まれている
ライセンスが設定されている
100文字以上の説明がある
GitHub Stars 100以上
3ヶ月以内に更新がある
10回以上フォークされている
オープンIssueが50未満
プログラミング言語が設定されている
1つ以上のタグが設定されている
Reviews
Reviews coming soon