c4-architecture
Skill0 starsChanged 10 days agoMITView on GitHub
Generate architecture documentation using C4 model Mermaid diagrams. Use when asked to create architecture diagrams, document system architecture, visualize software structure, create C4 diagrams, or generate context/container/component/deployment diagrams. Triggers include "architecture diagram", "C4 diagram", "system context", "container diagram", "component diagram", "deployment diagram", "document architecture", "visualize architecture".
14 other projects use a version of this, 8 of them edited.
Adopted by
Most recently taken up by
- selamy-labs/agent-skillstheir own version, first seen Nov 8, 2026
- delorenj/skillsedited, first seen Nov 7, 2026
- fsmiamoto/skillstheir own version, first seen Nov 1, 2026
- xiongxianfei/skillsmiththeir own version, first seen Nov 1, 2026
- coco-research/cocoedited, first seen Nov 1, 2026
- imMamdouhaboammar/kaku-chatgpt-harnessedited, first seen Oct 25, 2026
Other versions14 projects use a version of this.
Word for word · 0
None found.
Edited · 8
- coco-research/coco★ 79764% same
- delorenj/skills★ 1794% same
- eugenepyvovarov/mcpbundler-agent-skills-marketplace★ 1294% same
- shalomb/agent-skills★ 394% same
- rohinipachanta/claude-code-quickstart★ 181% same
- imMamdouhaboammar/kaku-chatgpt-harness★ 094% same
- nbirkner/digital-brain★ 094% same
- Samueladewole/claude-skills★ 094% same
Wrote their own version · 6
- microsoft/hve-core★ 1.6ksame name
- bitsmuggler/c4-skill★ 56same name
- sergeeey/Claude-cod-top-2026★ 9same name
- xiongxianfei/skillsmith★ 2same name
- fsmiamoto/skills★ 0same name
- selamy-labs/agent-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository20 other files it gives its agents.
- AGENTS.mdAGENTS.md
- examples/CLAUDE.mdCLAUDE.md
- api-and-interface-designSkill
- backlogSkill
- change-recordSkill
- ci-cd-and-automationSkill
- competitive-briefSkill
- documentation-and-adrsSkill
- frontend-designSkill
- frontend-ui-engineeringSkill
- git-workflow-and-versioningSkill
- oss-release-readinessSkill
- performance-optimizationSkill
- security-and-hardeningSkill
- shipping-and-launchSkill
- synthesize-researchSkill
- user-researchSkill
- ux-copySkill
- webapp-testingSkill
- web-visual-designSkill
The file
We split this skill into the rules it gives an agent
9 rules. 0 are common, 3 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Generate diagrams: - Create Mermaid C4 diagrams at appropriate abstraction levels
- Very specific: line 15
Document: - Write all diagrams, their descriptions and the AI YAML into one ARCHITECTURE.md (see Output Format)
- Very specific: line 16
Cross-check: - Every Mermaid alias has a YAML element with the same id, and every Rel has a YAML relation, and vice versa
- Very specific: line 195
Use $offsetX and $offsetY to fix overlapping relationship labels.
- Very specific: line 211
Meaningful aliases: - Use descriptive aliases (e.g., orderService not s1)
- Distinctive: line 212
Concise descriptions: - Keep descriptions under 50 characters when possible
- Very specific: line 275
Write one file, docs/architecture/ARCHITECTURE.md, readable by humans and by other people's AI agents:
- Distinctive: line 277
Humans: read the rendered Mermaid diagrams and the descriptions.
- Very specific: line 329
Every Rel in any diagram has a matching YAML relation (same from, to, label).
“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
C4 Architecture Documentation
Generate software architecture documentation using C4 model diagrams in Mermaid syntax.
Workflow
- Understand scope - Determine which C4 level(s) are needed based on audience
- Analyze codebase - Explore the system to identify components, containers, and relationships
- Generate diagrams - Create Mermaid C4 diagrams at appropriate abstraction levels
- Document - Write all diagrams, their descriptions and the AI YAML into one
ARCHITECTURE.md(see Output Format) - Cross-check - Every Mermaid alias has a YAML element with the same
id, and everyRelhas a YAML relation, and vice versa
C4 Diagram Levels
Select the appropriate level based on the documentation need:
| Level | Diagram Type | Audience | Shows | When to Create |
| 1 | C4Context | Everyone | System + external actors | Always (required) |
| 2 | C4Container | Technical | Apps, databases, services | Always (required) |
| 3 | C4Component | Developers | Internal components | Only if adds value |
| 4 | C4Deployment | DevOps | Infrastructure nodes | For production systems |
| - | C4Dynamic | Technical | Request flows (numbered) | For complex workflows |
Key Insight: "Context + Container diagrams are sufficient for most software development teams." Only create Component/Code diagrams when they genuinely add value.
Quick Start Examples
System Context (Level 1)
C4Context
title System Context - Workout Tracker
Person(user, "User", "Tracks workouts and exercises")
System(app, "Workout Tracker", "Vue PWA for tracking strength and CrossFit workouts")
System_Ext(browser, "Web Browser", "Stores data in IndexedDB")
Rel(user, app, "Uses")
Rel(app, browser, "Persists data to", "IndexedDB")Container Diagram (Level 2)
C4Container
title Container Diagram - Workout Tracker
Person(user, "User", "Tracks workouts")
Container_Boundary(app, "Workout Tracker PWA") {
Container(spa, "SPA", "Vue 3, TypeScript", "Single-page application")
Container(pinia, "State Management", "Pinia", "Manages application state")
ContainerDb(indexeddb, "IndexedDB", "Dexie", "Local workout storage")
}
Rel(user, spa, "Uses")
Rel(spa, pinia, "Reads/writes state")
Rel(pinia, indexeddb, "Persists", "Dexie ORM")Component Diagram (Level 3)
C4Component
title Component Diagram - Workout Feature
Container(views, "Views", "Vue Router pages")
Container_Boundary(workout, "Workout Feature") {
Component(useWorkout, "useWorkout", "Composable", "Workout execution state")
Component(useTimer, "useTimer", "Composable", "Timer state machine")
Component(workoutRepo, "WorkoutRepository", "Dexie", "Workout persistence")
}
Rel(views, useWorkout, "Uses")
Rel(useWorkout, useTimer, "Controls")
Rel(useWorkout, workoutRepo, "Saves to")Dynamic Diagram (Request Flow)
C4Dynamic
title Dynamic Diagram - User Sign In Flow
ContainerDb(db, "Database", "PostgreSQL", "User credentials")
Container(spa, "Single-Page App", "React", "Banking UI")
Container_Boundary(api, "API Application") {
Component(signIn, "Sign In Controller", "Express", "Auth endpoint")
Component(security, "Security Service", "JWT", "Validates credentials")
}
Rel(spa, signIn, "1. Submit credentials", "JSON/HTTPS")
Rel(signIn, security, "2. Validate")
Rel(security, db, "3. Query user", "SQL")
UpdateRelStyle(spa, signIn, $textColor="blue", $offsetY="-30")Deployment Diagram
C4Deployment
title Deployment Diagram - Production
Deployment_Node(browser, "Customer Browser", "Chrome/Firefox") {
Container(spa, "SPA", "React", "Web application")
}
Deployment_Node(aws, "AWS Cloud", "us-east-1") {
Deployment_Node(ecs, "ECS Cluster", "Fargate") {
Container(api, "API Service", "Node.js", "REST API")
}
Deployment_Node(rds, "RDS", "db.r5.large") {
ContainerDb(db, "Database", "PostgreSQL", "Application data")
}
}
Rel(spa, api, "API calls", "HTTPS")
Rel(api, db, "Reads/writes", "JDBC")Element Syntax
People and Systems
Person(alias, "Label", "Description")
Person_Ext(alias, "Label", "Description") # External person
System(alias, "Label", "Description")
System_Ext(alias, "Label", "Description") # External system
SystemDb(alias, "Label", "Description") # Database system
SystemQueue(alias, "Label", "Description") # Queue systemContainers
Container(alias, "Label", "Technology", "Description")
Container_Ext(alias, "Label", "Technology", "Description")
ContainerDb(alias, "Label", "Technology", "Description")
ContainerQueue(alias, "Label", "Technology", "Description")Components
Component(alias, "Label", "Technology", "Description")
Component_Ext(alias, "Label", "Technology", "Description")
ComponentDb(alias, "Label", "Technology", "Description")Boundaries
Enterprise_Boundary(alias, "Label") { ... }
System_Boundary(alias, "Label") { ... }
Container_Boundary(alias, "Label") { ... }
Boundary(alias, "Label", "type") { ... }Relationships
Rel(from, to, "Label")
Rel(from, to, "Label", "Technology")
BiRel(from, to, "Label") # Bidirectional
Rel_U(from, to, "Label") # Upward
Rel_D(from, to, "Label") # Downward
Rel_L(from, to, "Label") # Leftward
Rel_R(from, to, "Label") # RightwardDeployment Nodes
Deployment_Node(alias, "Label", "Type", "Description") { ... }
Node(alias, "Label", "Type", "Description") { ... } # ShorthandStyling and Layout
Layout Configuration
UpdateLayoutConfig($c4ShapeInRow="3", $c4BoundaryInRow="1")$c4ShapeInRow- Number of shapes per row (default: 4)$c4BoundaryInRow- Number of boundaries per row (default: 2)
Element Styling
UpdateElementStyle(alias, $fontColor="red", $bgColor="grey", $borderColor="red")Relationship Styling
UpdateRelStyle(from, to, $textColor="blue", $lineColor="blue", $offsetX="5", $offsetY="-10")Use $offsetX and $offsetY to fix overlapping relationship labels.
Best Practices
Essential Rules
- Every element must have: Name, Type, Technology (where applicable), and Description
- Use unidirectional arrows only - Bidirectional arrows create ambiguity
- Label arrows with action verbs - "Sends email using", "Reads from", not just "uses"
- Include technology labels - "JSON/HTTPS", "JDBC", "gRPC"
- Stay under 20 elements per diagram - Split complex systems into multiple diagrams
Clarity Guidelines
- Start at Level 1 - Context diagrams help frame the system scope
- One level per section - Each diagram stays on a single abstraction level, in its own section of
ARCHITECTURE.md - Meaningful aliases - Use descriptive aliases (e.g.,
orderServicenots1) - Concise descriptions - Keep descriptions under 50 characters when possible
- Always include a title - "System Context diagram for [System Name]"
What to Avoid
See references/common-mistakes.md for detailed anti-patterns:
- Confusing containers (deployable) vs components (non-deployable)
- Modeling shared libraries as containers
- Showing message brokers as single containers instead of individual topics
- Adding undefined abstraction levels like "subcomponents"
- Removing type labels to "simplify" diagrams
Microservices Guidelines
Single Team Ownership
Model each microservice as a container (or container group):
C4Container
title Microservices - Single Team
System_Boundary(platform, "E-commerce Platform") {
Container(orderApi, "Order Service", "Spring Boot", "Order processing")
ContainerDb(orderDb, "Order DB", "PostgreSQL", "Order data")
Container(inventoryApi, "Inventory Service", "Node.js", "Stock management")
ContainerDb(inventoryDb, "Inventory DB", "MongoDB", "Stock data")
}Multi-Team Ownership
Promote microservices to software systems when owned by separate teams:
C4Context
title Microservices - Multi-Team
Person(customer, "Customer", "Places orders")
System(orderSystem, "Order System", "Team Alpha")
System(inventorySystem, "Inventory System", "Team Beta")
System(paymentSystem, "Payment System", "Team Gamma")
Rel(customer, orderSystem, "Places orders")
Rel(orderSystem, inventorySystem, "Checks stock")
Rel(orderSystem, paymentSystem, "Processes payment")Event-Driven Architecture
Show individual topics/queues as containers, NOT a single "Kafka" box:
C4Container
title Event-Driven Architecture
Container(orderService, "Order Service", "Java", "Creates orders")
Container(stockService, "Stock Service", "Java", "Manages inventory")
ContainerQueue(orderTopic, "order.created", "Kafka", "Order events")
ContainerQueue(stockTopic, "stock.reserved", "Kafka", "Stock events")
Rel(orderService, orderTopic, "Publishes to")
Rel(stockService, orderTopic, "Subscribes to")
Rel(stockService, stockTopic, "Publishes to")
Rel(orderService, stockTopic, "Subscribes to")Output Format
Write one file, docs/architecture/ARCHITECTURE.md, readable by humans and by other people's AI agents:
- Humans read the rendered Mermaid diagrams and the descriptions.
- AI reads the Mermaid source and the YAML at the end. The YAML is the source of truth: if a diagram and the YAML disagree, the YAML wins.
Template (include only the levels the audience needs; Context and Container are required):
# <System> Architecture
## How to read this
Humans: diagrams and descriptions below. AI agents: the YAML in "AI structured data" lists every element and relation and is authoritative; `id` equals the Mermaid alias.
## 1. System Context
```mermaid
C4Context
title System Context - <System>
...
```
**Scope:** what this diagram covers and what it deliberately leaves out.
**Key decisions:** short bullets, link ADRs (`docs/decisions/001-....md`).
## 2. Container
```mermaid
C4Container
...
```
**Scope:** ...
**Key decisions:** ...
## 3. Component - <feature> (optional; also Deployment / Dynamic sections as needed)
## AI structured data
```yaml
elements:
- id: api # same as Mermaid alias, unique across all levels
level: container # context | container | component | deployment
type: container # person | system | system_ext | container | container_db | container_queue | component | node
name: API
tech: Node.js
description: REST API for orders
behavior: [Rejects orders without stock] # what it guarantees now
code: src/server/ # repo path; omit for external systems and people
parent: shop # optional: enclosing system/container/node id
relations:
- from: web
to: api
label: Creates orders
protocol: HTTPS/JSON
```Rules:
- One element appears once in the YAML even if it shows up in several diagrams; reuse the same alias in every diagram.
- Every
Relin any diagram has a matching YAML relation (samefrom,to,label); nothing in the YAML is missing from the diagrams. behaviorlists what the element guarantees now; update it when a change alters behavior.- Existing projects may list only the elements a change touches.
- Point the repo's
CLAUDE.md/AGENTS.mdat the file: "Architecture: readdocs/architecture/ARCHITECTURE.md."
Exporting Images (for Word or slides)
Render every diagram in the file to PNG next to it (ARCHITECTURE-1.png, ARCHITECTURE-2.png, ... in file order; a file with one diagram gives ARCHITECTURE.png):
bash scripts/render.sh <repo>/docs/architecture/ARCHITECTURE.mdThe script uses the Mermaid CLI pinned at ~/.local/mermaid-cli (11.17.0, with its own Chromium). Then open the PNG and look at it: Mermaid's C4 auto-layout often overlaps relationship labels. Fix with short labels and UpdateRelStyle(from, to, $offsetX=..., $offsetY=...) per relationship, re-render, and look again. On 11.17.0, UpdateLayoutConfig($c4ShapeInRow=...) did not change the column count.
Audience-Appropriate Detail
| Audience | Recommended Diagrams |
| Executives | System Context only |
| Product Managers | Context + Container |
| Architects | Context + Container + key Components |
| Developers | All levels as needed |
| DevOps | Container + Deployment |
References
- references/c4-syntax.md - Complete Mermaid C4 syntax
- references/common-mistakes.md - Anti-patterns to avoid
- references/advanced-patterns.md - Microservices, event-driven, deployment
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

