What is the Architecture Documentation Improvement Loop?

It is a specialized workflow that compares human-validated goal models with automatically generated code models to identify, document, and close architectural gaps.

How does the skill handle relationship parity?

It ensures consistency by construction, generating human-readable relationship tables and machine-usable LikeC4 code blocks simultaneously within the same Markdown source.

Can I use this with the LikeC4 VS Code extension?

Yes, the skill generates standard .c4 DSL files and a likec4.config.json, making it fully compatible with LikeC4 visualization tools.

What happens if my architecture changes?

You can use the 'devac architecture diff' and 'score' commands to identify exactly where the implementation has drifted from the validated documentation.

What are the M1-M4 quality rules?

These are meta-rules for documentation: M1 (Complete Sets), M2 (Generic Before Specific), M3 (State Recognition), and M4 (Handler Recognition).

Architecture Validator

Name: Architecture Validator
Author: pietgk

bypietgk

•

Aprendizaje y Documentación

Validates and maintains human-verified C4 architecture documentation using automated code-graph analysis and drift detection.

The Architecture Validator skill implements a robust documentation improvement loop by helping developers create, refine, and synchronize human-validated C4 models with actual code structure. By leveraging devac MCP tools to query code graphs, it identifies container groupings, key components, and data flows while applying rigorous M1-M4 documentation quality standards. This skill generates dual-format outputs—human-readable Markdown with ASCII diagrams and machine-usable LikeC4 DSL—allowing teams to detect architectural drift, calculate gap scores, and maintain a high-fidelity 'source of truth' for system design.

Características Principales

011 GitHub stars

02Automated package analysis via SQL code-graph queries and effect extraction

03Comprehensive audit trails through architectural reasoning files

04Relationship parity enforcement between visual diagrams and machine-readable tables

05Dual-format generation of Markdown documentation and LikeC4 DSL models

06Architectural drift detection with gap scoring and structural diffing

Casos de Uso

01Generating high-fidelity C4 diagrams for technical onboarding and stakeholder reviews

02Establishing a verified architectural baseline for legacy codebases or new projects

03Identifying and closing gaps between intended system design and actual implementation

Características Principales

011 GitHub stars

02Automated package analysis via SQL code-graph queries and effect extraction

03Comprehensive audit trails through architectural reasoning files

04Relationship parity enforcement between visual diagrams and machine-readable tables

05Dual-format generation of Markdown documentation and LikeC4 DSL models

06Architectural drift detection with gap scoring and structural diffing

Casos de Uso

01Generating high-fidelity C4 diagrams for technical onboarding and stakeholder reviews

02Establishing a verified architectural baseline for legacy codebases or new projects

03Identifying and closing gaps between intended system design and actual implementation