mirror of
https://github.com/NickCrew/Claude-Cortex.git
synced 2026-09-14 20:17:11 +08:00
15 KiB
15 KiB
Cortex Architecture Diagrams
Visual documentation of the three-layer automation system.
System Architecture Overview
graph TB
subgraph "Layer 1: User Interface"
User[👤 User Input]
Commands["/slash:commands<br/>43 commands<br/>16 namespaces"]
Flags["--flags<br/>--brainstorm, --ultrathink<br/>--focus, --orchestrate"]
end
subgraph "Layer 2: Behavioral Modes"
Brainstorm[🤝 Brainstorm<br/>Socratic dialogue]
DeepAnalysis[🔬 Deep Analysis<br/>Max reasoning]
QualityFocus[✨ Quality Focus<br/>8/10, 90%]
TokenEfficient[📦 Token Efficient<br/>30-50% reduction]
ParallelOrch[⚡ Parallel Orchestration<br/>Parallel execution]
SuperSaiyan[🔥 Super Saiyan<br/>Visual excellence]
TaskMgmt[📋 Task Management<br/>Hierarchical tasks]
end
subgraph "Layer 3: Workflows"
FeatureDev[🚀 Feature Development<br/>9 steps]
BugFix[🐛 Bug Fix<br/>8 steps]
Refactor[♻️ Refactoring<br/>10 steps]
APIDev[🎯 API Design<br/>15 steps]
TechDebt[🧹 Tech Debt<br/>13 steps]
ArchReview[🏗️ Architecture<br/>15 steps]
Onboard[👋 Onboarding<br/>17 steps]
Security[🔒 Security Audit<br/>8 steps]
Perf[⚡ Performance<br/>10 steps]
end
subgraph "Layer 4: Execution"
Agents[🤖 Specialized Agents<br/>code-reviewer, test-automator<br/>security-auditor, etc.]
MCP[🔌 MCP Servers<br/>Codanna, Context7<br/>Sequential]
Tools[🛠️ Native Tools<br/>Read, Write, Edit<br/>Bash, Grep, etc.]
end
User --> Commands
User --> Flags
Commands --> Brainstorm
Commands --> DeepAnalysis
Commands --> QualityFocus
Flags --> TokenEfficient
Flags --> ParallelOrch
Brainstorm --> FeatureDev
Brainstorm --> APIDev
Brainstorm --> Onboard
DeepAnalysis --> Refactor
DeepAnalysis --> TechDebt
DeepAnalysis --> ArchReview
QualityFocus --> Refactor
QualityFocus --> TechDebt
ParallelOrch --> FeatureDev
ParallelOrch --> Refactor
ParallelOrch --> APIDev
FeatureDev --> Agents
BugFix --> Agents
Refactor --> Agents
APIDev --> Agents
Agents --> MCP
Agents --> Tools
style User fill:#e1f5ff
style Commands fill:#ffe1e1
style Brainstorm fill:#fff3e1
style DeepAnalysis fill:#fff3e1
style QualityFocus fill:#fff3e1
style FeatureDev fill:#e8f5e9
style Agents fill:#f3e5f5
style MCP fill:#f3e5f5
Command → Mode → Workflow Flow
sequenceDiagram
participant U as 👤 User
participant C as /slash:command
participant M as 🎨 Mode
participant W as 🔄 Workflow
participant A as 🤖 Agents
participant T as 📋 TodoWrite
U->>C: /refactor:analyze src/auth
Note over C: Parse command<br/>Load configuration
C->>M: Activate Deep_Analysis
C->>M: Activate Quality_Focus
Note over M: Set behavior:<br/>- Max reasoning depth<br/>- Quality threshold 8/10<br/>- Coverage target 90%
M->>W: Trigger refactoring workflow
Note over W: Load workflow steps<br/>Initialize state
W->>T: Create progress tracking
T-->>U: Show task list
W->>A: Execute Step 1: Code Analysis
A->>A: code-reviewer agent
A-->>W: quality_report, refactoring_candidates
W->>T: Update: Step 1 ✅
T-->>U: Progress: 1/10 (10%)
W->>A: Execute Step 2: Impact Assessment
A->>A: Codanna MCP (analyze_impact)
A-->>W: dependency_graph, risk_assessment
W->>T: Update: Step 2 ✅
T-->>U: Progress: 2/10 (20%)
W->>A: Execute Step 3: Refactoring Plan
A->>A: Planning with Deep_Analysis
A-->>W: refactoring_plan, safety_checkpoints
W->>T: Update: Step 3 ✅
Note over W: Continue through<br/>remaining steps...
W-->>U: ✅ Workflow Complete<br/>📄 Refactoring plan generated
Mode Activation Flow
flowchart TB
Start([User Input]) --> Detect{Detect Trigger?}
Detect -->|Manual| Manual["/mode:activate"]
Detect -->|Flag| Flag["--brainstorm<br/>--ultrathink<br/>--focus quality"]
Detect -->|Auto| Auto["Keywords:<br/>'maybe', 'not sure'<br/>'refactor', 'analyze'"]
Detect -->|Workflow| Workflow["Workflow step<br/>specifies mode"]
Manual --> LoadMode[Load Mode Configuration]
Flag --> LoadMode
Auto --> LoadMode
Workflow --> LoadMode
LoadMode --> ParseDef[Parse Mode Definition]
ParseDef --> SetBehavior[Set Behavioral Changes]
SetBehavior --> Behavior{Mode Type?}
Behavior -->|Brainstorm| B1["- Ask clarifying questions<br/>- Present options<br/>- Use AskUserQuestion<br/>- Build shared understanding"]
Behavior -->|Deep Analysis| B2["- Extended reasoning depth<br/>- Hypothesis generation<br/>- Evidence-based conclusions<br/>- Transparency markers"]
Behavior -->|Quality Focus| B3["- Review threshold: 8/10<br/>- Coverage target: 90%<br/>- Elevated standards<br/>- Zero tolerance issues"]
Behavior -->|Token Efficient| B4["- Symbol communication<br/>- 30-50% reduction<br/>- Eliminate redundancy<br/>- Structured compression"]
Behavior -->|Parallel Orch| B5["- Parallel-first mindset<br/>- Quality gates mandatory<br/>- Agent maximization<br/>- Workstream coordination"]
B1 --> Active[Mode Active ✅]
B2 --> Active
B3 --> Active
B4 --> Active
B5 --> Active
Active --> Execute[Execute Task<br/>with Mode Behavior]
Execute --> Deactivate{Deactivate?}
Deactivate -->|Session End| End1([Mode Ends])
Deactivate -->|Task Complete| End2([Mode Ends])
Deactivate -->|Manual /mode:deactivate| End3([Mode Ends])
Deactivate -->|Permanent| Persist[Persist to Profile]
Persist --> End4([Mode Remains Active])
style Start fill:#e1f5ff
style LoadMode fill:#fff3e1
style Active fill:#c8e6c9
style Execute fill:#bbdefb
Workflow Execution Flow
flowchart TD
Start(["/workflow:run feature-development"]) --> Load[Load Workflow YAML]
Load --> Validate{Validate<br/>Definition?}
Validate -->|Invalid| Error1[❌ Report Error]
Validate -->|Valid| Init[Initialize Workflow State]
Init --> CreateTasks[Create TodoWrite Tasks<br/>for Each Step]
CreateTasks --> SetupOutputs[Setup Output Directory<br/>./workflow-outputs/]
SetupOutputs --> Step1[Step 1: Architecture Design]
Step1 --> LoadMode1[Load Required Mode:<br/>Deep_Analysis]
LoadMode1 --> Agent1[Activate Agent:<br/>cloud-architect]
Agent1 --> Execute1[Execute Step 1]
Execute1 --> Validate1{Outputs<br/>Valid?}
Validate1 -->|No| Retry1[Retry or Fail]
Validate1 -->|Yes| Update1[Update TodoWrite: ✅]
Update1 --> Checkpoint1[Save Checkpoint]
Checkpoint1 --> Step2[Step 2: Implementation Planning]
Step2 --> Agent2[Activate Agent:<br/>orchestrator]
Agent2 --> Execute2[Execute Step 2<br/>with Step 1 Outputs]
Execute2 --> Update2[Update TodoWrite: ✅]
Update2 --> Step3[Step 3: Code Implementation]
Step3 --> LoadMode3[Load Mode:<br/>Parallel_Orchestration]
LoadMode3 --> Parallel{Parallel<br/>Capable?}
Parallel -->|Yes| Multi[Launch Multiple Agents<br/>in Parallel]
Parallel -->|No| Single[Sequential Execution]
Multi --> Sync[Synchronize Results]
Single --> Sync
Sync --> Update3[Update TodoWrite: ✅]
Update3 --> Step4[Step 4: Code Review]
Step4 --> QualityGate{Quality Gate}
QualityGate -->|Fail| Block["❌ Block Progression<br/>Report Issues<br/>Request Fixes"]
QualityGate -->|Pass| Update4[Update TodoWrite: ✅]
Block --> Remediate[Fix Issues]
Remediate --> QualityGate
Update4 --> MoreSteps{More<br/>Steps?}
MoreSteps -->|Yes| NextStep[Continue to Next Step]
MoreSteps -->|No| CheckSuccess{All Success<br/>Criteria Met?}
NextStep --> Step1
CheckSuccess -->|No| Incomplete[❌ Workflow Incomplete<br/>Report Missing Criteria]
CheckSuccess -->|Yes| Report[Generate Final Report]
Report --> Archive[Archive Outputs]
Archive --> Complete([✅ Workflow Complete])
style Start fill:#e1f5ff
style Complete fill:#c8e6c9
style QualityGate fill:#fff3e1
style Block fill:#ffcdd2
style Archive fill:#c8e6c9
Refactoring Example: End-to-End
graph TB
subgraph "User Action"
U1["User: /refactor:analyze src/auth"]
end
subgraph "Command Layer"
C1["Command: refactor:analyze"]
C2["Parse arguments:<br/>path = 'src/auth'<br/>focus = 'all'"]
end
subgraph "Mode Activation"
M1["Activate: Deep_Analysis<br/>- Reasoning depth: ~32K tokens<br/>- Transparency markers on"]
M2["Activate: Quality_Focus<br/>- Review threshold: 8/10<br/>- Coverage target: 90%"]
end
subgraph "Workflow Execution"
W1["Load: refactoring.yaml"]
W2["Step 1: Code Analysis<br/>Agent: code-reviewer"]
W3["Step 2: Impact Assessment<br/>MCP: Codanna analyze_impact"]
W4["Step 3: Refactoring Plan<br/>Mode: Deep_Analysis"]
end
subgraph "Agent Layer"
A1["code-reviewer:<br/>- Complexity: 12 → target 6<br/>- Duplication: 45% → target 15%<br/>- Code smells: 8 found"]
A2["Codanna MCP:<br/>- Dependencies: 23 files<br/>- Risk score: 6.5/10<br/>- Change radius: Medium"]
A3["Planning:<br/>- Priority matrix generated<br/>- 3 quick wins identified<br/>- Migration strategy defined"]
end
subgraph "Output"
O1["📄 Refactoring Plan:<br/>- Issues prioritized<br/>- Impact assessed<br/>- Recommendations clear<br/>- Safety validated"]
end
subgraph "Next Action"
N1["User: /refactor:execute plan.md"]
N2["Command: refactor:execute<br/>Mode: Parallel_Orchestration"]
N3["Incremental Execution:<br/>- Change 1 → Test → ✅ Commit<br/>- Change 2 → Test → ✅ Commit<br/>- Change 3 → Test → ✅ Commit"]
N4["Quality Gates:<br/>- Review: 8.5/10 ✅<br/>- Coverage: 87% ✅<br/>- Performance: +2% ✅"]
N5["✅ Refactoring Complete"]
end
U1 --> C1
C1 --> C2
C2 --> M1
C2 --> M2
M1 --> W1
M2 --> W1
W1 --> W2
W2 --> W3
W3 --> W4
W2 --> A1
W3 --> A2
W4 --> A3
A1 --> O1
A2 --> O1
A3 --> O1
O1 --> N1
N1 --> N2
N2 --> N3
N3 --> N4
N4 --> N5
style U1 fill:#e1f5ff
style M1 fill:#fff3e1
style M2 fill:#fff3e1
style O1 fill:#c8e6c9
style N5 fill:#c8e6c9
Decision Tree: Which Layer to Use?
flowchart TD
Start{What do you<br/>need to do?}
Start -->|Quick action| Q1{Specific<br/>task?}
Start -->|Change behavior| Q2{How Claude<br/>operates?}
Start -->|Multi-step process| Q3{Complex<br/>workflow?}
Q1 -->|Yes| UseCommand["✅ Use Slash Command<br/>Example:<br/>/refactor:analyze<br/>/test:generate-tests<br/>/docs:generate"]
Q2 -->|Reasoning depth| UseMode1["✅ Use Mode<br/>Deep_Analysis<br/>--ultrathink"]
Q2 -->|Quality standards| UseMode2["✅ Use Mode<br/>Quality_Focus<br/>--focus quality"]
Q2 -->|Communication style| UseMode3["✅ Use Mode<br/>Token_Efficient<br/>--uc"]
Q2 -->|Execution pattern| UseMode4["✅ Use Mode<br/>Parallel_Orchestration<br/>--orchestrate"]
Q3 -->|Feature development| UseWorkflow1["✅ Use Workflow<br/>/workflow:run<br/>feature-development"]
Q3 -->|Code refactoring| UseWorkflow2["✅ Use Workflow<br/>/workflow:run<br/>refactoring"]
Q3 -->|API design| UseWorkflow3["✅ Use Workflow<br/>/workflow:run<br/>api-design"]
Q3 -->|Developer onboarding| UseWorkflow4["✅ Use Workflow<br/>/workflow:run<br/>onboarding"]
UseCommand --> Examples1["Commands are:<br/>✓ Fast<br/>✓ Focused<br/>✓ Single purpose"]
UseMode1 --> Examples2["Modes are:<br/>✓ Behavioral<br/>✓ Persistent<br/>✓ Context-aware"]
UseMode2 --> Examples2
UseMode3 --> Examples2
UseMode4 --> Examples2
UseWorkflow1 --> Examples3["Workflows are:<br/>✓ Multi-step<br/>✓ Coordinated<br/>✓ Progressive"]
UseWorkflow2 --> Examples3
UseWorkflow3 --> Examples3
UseWorkflow4 --> Examples3
Examples1 --> Combine{Can be<br/>combined?}
Examples2 --> Combine
Examples3 --> Combine
Combine -->|Yes| Combined["✅ Example:<br/>/workflow:run refactoring<br/>└─ Activates: Deep_Analysis<br/>└─ Activates: Quality_Focus<br/>└─ Executes: 10 workflow steps<br/>└─ Uses: code-reviewer agent"]
style Start fill:#e1f5ff
style UseCommand fill:#ffe1e1
style UseMode1 fill:#fff3e1
style UseMode2 fill:#fff3e1
style UseMode3 fill:#fff3e1
style UseMode4 fill:#fff3e1
style UseWorkflow1 fill:#e8f5e9
style UseWorkflow2 fill:#e8f5e9
style UseWorkflow3 fill:#e8f5e9
style UseWorkflow4 fill:#e8f5e9
style Combined fill:#c8e6c9
System Statistics
pie title System Components Distribution
"Slash Commands" : 43
"Modes" : 8
"Workflows" : 9
"Agents" : 29
"MCP Servers" : 3
graph LR
subgraph "Coverage"
A[16 Command Namespaces]
B[8 Behavioral Modes]
C[9 Workflow Types]
D[29 Specialized Agents]
E[3 MCP Servers]
end
style A fill:#ffe1e1
style B fill:#fff3e1
style C fill:#e8f5e9
style D fill:#f3e5f5
style E fill:#e1f5ff
Mode + Workflow Compatibility Matrix
graph TB
subgraph "Workflows"
W1[Feature Dev]
W2[Bug Fix]
W3[Refactoring]
W4[API Design]
W5[Tech Debt]
W6[Arch Review]
W7[Onboarding]
W8[Security]
W9[Performance]
end
subgraph "Modes"
M1[Brainstorm]
M2[Deep Analysis]
M3[Quality Focus]
M4[Parallel Orch]
M5[Task Mgmt]
end
M1 -.-> W1
M1 -.-> W4
M1 -.-> W7
M2 -.-> W3
M2 -.-> W5
M2 -.-> W6
M3 -.-> W3
M3 -.-> W5
M3 -.-> W8
M4 -.-> W1
M4 -.-> W3
M4 -.-> W4
M5 -.-> W1
M5 -.-> W7
style W1 fill:#e8f5e9
style W2 fill:#e8f5e9
style W3 fill:#e8f5e9
style W4 fill:#e8f5e9
style W5 fill:#e8f5e9
style W6 fill:#e8f5e9
style W7 fill:#e8f5e9
style W8 fill:#e8f5e9
style W9 fill:#e8f5e9
style M1 fill:#fff3e1
style M2 fill:#fff3e1
style M3 fill:#fff3e1
style M4 fill:#fff3e1
style M5 fill:#fff3e1
Legend
| Symbol | Meaning |
|---|---|
| 👤 | User / Human Input |
| 🎨 | Behavioral Mode |
| 🔄 | Workflow / Process |
| 🤖 | Agent / AI Actor |
| 🔌 | MCP Server |
| 🛠️ | Native Tool |
| ✅ | Success / Complete |
| ❌ | Failure / Error |
| 📋 | Task / Todo |
| 📄 | Output / Document |
| ⚡ | Fast / Optimized |
| 🔒 | Security |
| 🎯 | Target / Goal |
Quick Start Guide
For Developers
-
Explore the system:
cortex # Press 3 → View Modes # Press 6 → View Workflows -
Try a command:
/refactor:analyze src/ /workflow:run feature-development /mode:activate Brainstorm -
Understand the flow:
- Command activates → Mode influences → Workflow executes → Agents work
For Architects
- Review diagrams in this file
- Understand three layers: Commands → Modes → Workflows
- Study integration patterns: How layers interact
- Customize workflows: Adapt to your needs
For DevOps
- Deployment workflows:
feature-development,security-audit - Quality gates: Integrated in all workflows
- Automation ready: Commands can be scripted
- State management: Checkpoint/resume support