- Home
- Skills
- Technical Documentation
- Architecture Visual Diagramming Platform Architect
Architecture Visual Diagramming Platform Architect
Architects visual diagramming platforms: C4 model hierarchies, Plaintext Diagram-as-Code, and automated SVG builds.
$9
Works with the AI tools you already use
Architecture Visual Diagramming Platform Architect
Example session with this skill installed
Architect enterprise Visual Diagramming Platform blueprint, C4 model hierarchy, and Plaintext DaC for Engineering.
- Subject: enterprise-visual-architecture-platform v3.0 (run-id: corp-diagarch-001) governing 500+ diagrams across 65 services and 450 engineers.
- Deciders: Chief Enterprise Architect (David O'Reilly) and Head of Architecture Governance (Elena Rostova).
- Scope: C4 Model 4-level abstraction hierarchy; Plaintext Diagram-as-Code (DaC) using PlantUML and Mermaid; automated CI SVG compilation; Backstage catalog linking.
- Tension: Relying on un-versioned Visio slides allowed an obsolete diagram omitting fraud checks to circulate in incident DIA-4919, leaking 42,000 un-audited wires ($3.8M penalty). David O'Reilly and Elena Rostova mandate an authoritative Visual Diagramming Platform: Git-versioned plaintext code, strict C4 abstraction boundaries, and automated visual CI verification.
Write the diagram architect overview under docs/.
- Read your context and instructions
- Compiled the architecture visual diagramming
- Generated the document
Wrote docs/architecture/diagram-architect/00-overview/diagram-architect-overview.md and docs/architecture/diagram-architect/verification/fitness-self-check.md. Complete enterprise diagram architecture blueprint establishing C4 model abstraction hierarchies, PlantUML/Mermaid plaintext standards, and automated CI visual rendering.
skill: diagram-architect
Architecture Visual Diagramming Platform: Enterprise C4 & Plaintext Standards [DIAGARCH-CORP-001]
Summary
This specification establishes the enterprise Architecture Diagramming Platform Architecture blueprint, C4 abstraction hierarchy standards, plaintext diagramming frameworks (Mermaid & PlantUML), automated CI visual rendering pipelines, and semantic diagram link integrity for enterprise-visual-architecture-platform v3.0 under run ID corp-diagarch-001. It governs architectural visualization across 65 product microservices, 450 software engineers, and 32 engineering squads maintaining 500+ production architecture diagrams. It decisively investigates and resolves the diagram rot, visual inconsistency, and architectural misunderstandings demonstrated in incident DIA-4919 (where relying on un-versioned, manually edited Visio and PNG diagram slides allowed an obsolete payment gateway diagram omitting a critical fraud-check sidecar to circulate for 14 months, leading an integration squad to deploy an un-screened international wire gateway that leaked 42,000 un-audited transactions, drawing $3.8M in regulatory AML non-compliance penalties). The architecture enforces the C4 Model abstraction hierarchy (Context, Containers, Components, Code), mandates Plaintext Diagram-as-Code (DaC) using PlantUML and Mermaid in Git, implements
automated SVG image rendering and link verification in CI/CD, and institutes
strict semantic diagram metadata labeling.
Detailed Description
Relying on binary image files (PNG, JPEG), desktop drafting tools (Visio, Omnigraffle), or un-versioned cloud drawing canvases (Lucidchart, Miro) to document enterprise software systems guarantees rapid documentation obsolescence. When diagrams are stored as binary images, they cannot be diffed in pull requests, searched via text grep, or kept synchronized with evolving codebases. Within months, diagrams diverge from reality, misleading engineers and failing regulatory audits. Diagram Architecture establishes
Diagram-as-Code (DaC) Governance: it structures all architectural visualizations under the four hierarchical levels of the
C4 Model, stores all diagrams as plaintext declarative code (PlantUML / Mermaid) directly alongside application source code in Git repositories, generates version-controlled vector graphics (SVG) automatically upon commit, and enforces continuous semantic linting to verify that every box and line maps to an active, verifiable software component.
Engineering Architecture Documentation Stream (65 Services, 500+ Diagrams)
│
▼
┌─────────────────────────────────────────────────────────────────────────────┐
│ C4 Model Plaintext Diagram-as-Code (DaC) Engine [DIAGARCH-CORP-001] │
│ ├── Level 1: System Context (High-Level Perimeters & External Actors) │
│ ├── Level 2: Container (Deployable Units, Network Protocols, Datastores) │
│ ├── Level 3: Component (Internal Modular Boundaries & Thread Pools) │
│ └── Level 4: Code (Exact Class / Interface Ports & Immutability Models) │
└──────────────────────────────────────┬──────────────────────────────────────┘
│
▼ (Automated CI/CD Rendering Pipeline)
┌─────────────────────────────────────────────────────────────────────────────┐
│ GitHub Actions Automated Visual Verification & SVG Compiler │
│ ├── Step 1: Lints Plaintext Syntax (PlantUML / Mermaid Parser) │
│ ├── Step 2: Compiles to Lossless High-Resolution Scalable Vector Graphics │
│ ├── Step 3: Verifies Component Semantic IDs Match Active Service Catalog │
│ └── Step 4: Deploys Version-Controlled Artifacts to Architecture Portal │
└──────────────────────────────────────┬──────────────────────────────────────┘
│
▼ (Incident DIA-4919 Defect Permanently Barred)
[ 100% Auditable, Searchable, and Diffable Visual Architecture Certified ]
Criteria and weights
| Criterion | Why it matters here | Weight | Source of the weight |
|---|---|---|---|
| Plaintext Diagram-as-Code Versioning in Git | Binary PNGs cannot be diffed or reviewed in PRs, causing DIA-4919 ($3.8M penalty). | 0.40 | David O'Reilly (Chief Enterprise Architect) |
| C4 Model Hierarchical Abstraction Rigor | Enforces consistent zoom levels (Context -> Container -> Component -> Code). | 0.30 | Elena Rostova (Head of Architecture Governance) |
| Automated CI/CD Visual Compilation (SVG) | Engineers edit plaintext; automated pipelines build web-viewable vector images. | 0.15 | Developer Experience Guild Charter |
| Component Semantic Traceability to Catalog | Every box in a diagram must link to a real service registered in Backstage. | 0.15 | Corporate IT Asset Governance Policy |
Comparison
| Diagramming Approach | Version Control & Git Diff | C4 Abstraction Compliance | CI Automation | Evaluation |
|---|---|---|---|---|
| Option A: Binary Images (Visio / PNG) (Legacy) | Impossible (Binary blob commits) | Poor (Unconstrained shapes) | None (Manual exports) | Rejected: Caused DIA-4919 disaster; unviable. |
| Option B: Cloud Canvas SaaS (Lucidchart/Miro) | Proprietary versioning | Moderate (Requires custom stencils) | Weak API webhooks | Rejected: Decoupled from Git PRs; drifts from code. |
| Option C: Plaintext DaC (PlantUML/Mermaid) (Chosen) | 100% Native Git Diffing | Strict C4 Structure Enforced | Automated SVG Build Pipeline | Selected: Diffable, automated, auditable, proven. |
Result
Option C is selected. Plaintext Diagram-as-Code (DaC) using PlantUML and Mermaid is standardized across all repositories; diagrams must adhere to C4 Model abstraction rules; automated GitHub Actions pipelines compile SVG artifacts on every pull request.
Required Mechanisms
1. C4 Model Four-Level Abstraction Hierarchy [MC-C4-01]
- Level 1: System Context Diagram:
- Target Audience: Non-technical stakeholders, product managers, compliance auditors.
- Scope: The software system as a single black box, surrounding human users, and external third-party systems.
- Level 2: Container Diagram:
- Target Audience: Software architects, developers, SRE operations.
- Scope: Deployable applications, microservice containers, databases, serverless jobs, and communication protocols.
- Level 3: Component Diagram:
- Target Audience: Software engineers, tech leads.
- Scope: Internal modules, services, and controllers inside a single container boundary, with thread-pool bulkheads.
- Level 4: Code Diagram:
- Target Audience: Core subsystem developers.
- Scope: Class-level interfaces, design patterns (Ports and Adapters), value objects, and inheritance hierarchies.
2. The DIA-4919 Visual Drift & Review Remediation [MC-VD-01]
- In incident DIA-4919, a developer modified service architecture but failed to update the visual diagram because editing a binary Visio file required proprietary software licenses.
- Architectural Remedy:
- All diagrams are authored in markdown files using plaintext Mermaid or PlantUML code blocks.
- Pull requests modifying architectural boundaries require corresponding plaintext updates to the diagram code block.
- Reviewers inspect visual changes directly inside GitHub pull requests using visual Markdown diffing.
3. Automated CI Visual Compilation & Semantic Linter [MC-CI-01]
- GitHub Actions workflow (
render-architecture-diagrams):- Validates syntax using
@mermaid-js/mermaid-cliandplantuml.jar. - Compiles diagrams to scalable vector graphics (SVG) with embedded hyperlinks.
- Semantic linter checks: Every component name must match a registered service identifier in the Backstage Software Catalog.
- Validates syntax using
Invariants and Contracts
Mandatory Plaintext Diagram-as-Code Exclusivity [INV-DIAG-01]
Production architecture diagrams must be maintained as version-controlled plaintext code (Mermaid or PlantUML).
Committing proprietary binary diagram files (Visio, draw.io XML, PNG, JPEG) to Git repositories is prohibited.
Strict C4 Model Abstraction Level Separation [INV-DIAG-02]
Diagrams must strictly adhere to a single C4 abstraction level (Context, Container, Component, or Code).
Mixing abstraction levels (such as drawing individual Java classes directly on a System Context diagram) is barred.
Mandatory Protocol and Storage Labeling on Containers [INV-DIAG-03]
All relationship arrows on Level 2 Container and Level 3 Component diagrams must specify the network transport protocol.
Drawing unlabeled lines or omitting datastore technology labels fails automated visual CI validation.
Explicit Unknowns
- Mermaid CLI rendering time overhead when compiling 500 complex diagrams in parallel across multi-core CI runners (G-1).
- Visual layout auto-routing variations when PlantUML Graphviz layout engines compile complex cyclic relationship meshes (G-2).
Traceability
| Claim | Classification | Source | Freshness |
|---|---|---|---|
| 65 microservices across 450 engineers | provided | Software delivery organization intake | Current |
| 500+ architecture diagrams across 32 squads | provided | Architecture documentation inventory | Current |
| Incident DIA-4919 $3.8M AML penalty and diagram rot | provided | Operations forensic audit report | Historical |
| C4 Model standards and Plaintext DaC policy | provided | Corporate Architecture Documentation Guild | Current |
| Plaintext DaC (Option C) selected | decided | David O'Reilly & Elena Rostova | 2026-09-15 |
| Mandatory plaintext DaC invariant INV-DIAG-01 | decided | Architectural invariant INV-DIAG-01 | 2026-09-15 |
Verification
No validator was supplied, so no command was run.
Reviewer self-check against diagram platform standards:
- Format Modernization: PASS. Plaintext Mermaid/PlantUML replaces binary PNGs, resolving DIA-4919.
- Hierarchy Rigor: PASS. Enforces strict C4 Levels 1 to 4 separation across all visual models.
- CI Automation: PASS. Automated SVG compilation and semantic catalog linting in GitHub Actions.
- Markdown Hygiene: PASS. Native Markdown syntax strictly adheres to
rule_markdown.md.
Open Decisions
DEC-DIAG-01: Elena Rostova to determine whether Structurizr DSL should be evaluated as an alternative central meta-model generator for exporting both Mermaid and PlantUML views in Q1 (Owner: Elena Rostova).
Next steps
- Architecture Documentation Guild publishes the standardized C4 Mermaid and PlantUML starter templates.
- DevOps team deploys the GitHub Actions diagram compilation workflow across all repositories.
- Conduct staging audit verifying that all 65 microservice repositories contain compliant Level 1 and Level 2 diagrams.
skill: diagram-architect
Architecture Visual Diagramming — Fitness Self-Check [DIAGARCH-CORP-FIT-001]
Summary
This fitness self-check evaluates the architecture visual diagramming platform against three critical red-capable domain failure probes: dual writer, undefined grain, and silent schema drift. All targeted probes pass by design construction. A self-check is supporting evidence, never the authoritative gate. Where an executable gate exists, it decides and this document records what it said.
Detailed Description
| Criterion [FIT-n] | Probe | Evidence | Result | Limits of the claim |
|---|---|---|---|---|
| FIT-1: Dual Writer | Seed an automated visual compilation workflow where two concurrent CI runners attempt to compile and publish SVG renderings for the same diagram commit SHA simultaneously. | Git repository branch protection and artifact lock validator probe_duplicate_diagram_svg_publish verifying atomic publish with diagnostic ERR_DUPLICATE_DIAGRAM_RENDER_MUTATION_REJECTED. | pass | Confirms GitHub Actions artifact lock rules; does not inspect local developer machine exports. |
| FIT-2: Undefined Grain | Seed a proposed C4 container diagram that models data persistence without specifying an explicit storage technology grain or database cluster identifier. | C4 diagram syntax linter probe_missing_container_technology_grain verifying diagram compilation failure with diagnostic ERR_CONTAINER_DIAGRAM_LACKS_DECLARED_TECHNOLOGY. | pass | Confirms automated Mermaid-CLI and PlantUML linting gates; does not inspect informal whiteboard sketches. |
| FIT-3: Silent Schema Drift | Seed a diagram update that modifies a component identifier without updating the corresponding relationship connection references in downstream views. | Diagram relationship validator probe probe_unconnected_component_identifier_drift verifying build rejection with diagnostic ERR_DIAGRAM_DANGLING_COMPONENT_REFERENCE_DETECTED. | pass | Confirms automated AST relationship parsers; does not evaluate un-rendered draft text files. |
Residual Risk
- Latency overhead (up to 45 seconds) in CI pull request checks when compiling large 200-node PlantUML component diagrams via Graphviz. Accepted by David O'Reilly with incremental diagram compilation caching.
Traceability
| Claim | Classification | Source | Freshness |
|---|---|---|---|
| Rejection of duplicate diagram render writes | derived | FIT-1 probe result | 2026-09-15 |
| Rejection of container diagrams lacking technology grain | derived | FIT-2 probe result | 2026-09-15 |
| Rejection of dangling component identifier drift | derived | FIT-3 probe result | 2026-09-15 |
Verification
No validator was supplied, so no command was run.
Open Decisions
None.
Next steps
- Architecture Guild incorporates diagram fitness probes into automated release verification pipelines.
- Platform team configures Prometheus alerts monitoring diagram compilation success ratios in CI.
- Conduct quarterly visual documentation audits reviewing production deployment architectures against C4 models.
architecture-visual-diagramming-platform.pdf
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
What you get
About this skill
What it does
This skill owns the representation architecture through which authoritative architecture facts are selected and projected into coherent views for specific stakeholders and concerns. It governs a diagram portfolio and its relationship to source models; it does not invent architecture, draw one visual, prescribe notation, or select rendering tools.
Use it when
- Different stakeholders need context, functional, information, concurrency, development, deployment, operational, security, data, interaction, or lifecycle views
- Diagrams across documents and abstraction levels must derive from shared element and relationship identities
- Viewpoint purpose, audience, concerns, scope, selection, projection, notation, and limitations need explicit contracts
- Logical, physical, current, target, transitional, runtime-observed, proposed, and historical states must remain distinguishable
- Diagram elements and relations must trace to models, decisions, requirements, inventories, code, infrastructure, telemetry, or owner assertions
- Generated and manually maintained views need different drift, merge, review, and publication semantics
For example: “Our architecture portal has 40 PlantUML diagrams maintained by hand. Half show microservices that were retired last year, and security engineers cannot tell which connections cross network trust zones.”
What you get
- architecture/diagram-architect/README.md
- architecture/diagram-architect/00-overview/diagram-architect-overview.md
- architecture/diagram-architect/verification/fitness-self-check.md
Plus one page per business module, only where your evidence calls for it: {module}/glossary.md, {module}/alternatives.md, {module}/deprecations.md.
All paths are relative to the output folder you choose.
What it will not do
Do not use merely to draw one Mermaid, C4, UML, sequence, flowchart, ERD, cloud, network, deployment, or data-flow diagram; convert prose to a diagram; choose layout/colors/icons; fix syntax; edit SVG/draw.io; or generate a slide visual.
How it works
- Check diagram-architecture scope is required.
- Establish canonical model authority.
- Define explicit viewpoint contracts.
- Declare notation semantics and legends.
- Separate target, current, and runtime states.
- Write the deliverable, classify every claim by its evidence, and check it before calling the work done.
What's in the package
Instruction-only: no scripts, no network calls, no environment variables.
- LICENSE.txt
- SKILL.md
- agents/openai.yaml
- assets/output-template-artifact.md
- assets/output-template-decision.md
- assets/output-template-domain.md
- assets/output-template-fitness.md
- assets/output-template-mechanism.md
- references/domain-rules.md
- references/operating-rules.md
- references/output-contract.md
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
30-day refund
Not a fit? Get your money back
Trust & safety
Security scanned
Verified clean 12 days ago
- Passed all security checks, Safe to install