AgenticForce Orchestrator

Enterprise Salesforce development template optimized for AI-native workflows with Claude, Cursor, and Cline integration.

Project Context & Purpose

This project serves as a comprehensive template for modern, AI-assisted Salesforce development, demonstrating enterprise-grade practices while maintaining beginner accessibility.

Key Differentiators:

Dual-audience documentation (human vs AI-specific context)
Pre-configured AI development environment with .cursor/ and memory-bank/
Realistic sample data and comprehensive test coverage
Automated CI/CD validation via GitHub Actions

AI Assistant Guidelines

Essential Context Files (Always Reference These)

Enterprise Patterns:
```
memory-bank/enterprise-patterns.md
```
Integration Guidelines:
```
memory-bank/integration-patterns.md
```
Project Overview:
```
memory-bank/project-overview.md
```
AI Rules:
```
.cursor/rules/core.mdc
```
Memory Bank Index:
```
.cursor/rules/memory-bank.mdc
```

When Working on This Project:

Always reference memory-bank/ files for established patterns
Follow .cursor/rules/ for project-specific guidelines
Use existing sample data patterns in data/ folder
Maintain test coverage standards (current: 25KB+ test classes)

Coding Standards:

Follow Salesforce naming conventions and bulkification patterns
Implement comprehensive test coverage (target: >85%)
Use descriptive commit messages following conventional commits
Reference existing architectural patterns before creating new ones

Architecture Overview

Project Structure

├── .cursor/rules/           # AI coding guidelines and context
├── memory-bank/            # Persistent AI knowledge base  
├── force-app/main/default/ # Salesforce metadata
├── docs/onboarding/        # Human-facing documentation
├── data/                   # Sample CSV data for testing
├── scripts/deploy/         # Deployment automation
└── .github/workflows/      # CI/CD validation (now working!)

Salesforce Components

Custom Objects: Project__c with validation rules and automation
Apex Classes: HelloWorld (13KB), TestDataFactory, LoadDataTestExample (6KB)
LWC Components: projectCard for modern, responsive UI
Automation: Flag_Overdue_Projects flow for business logic
Security: ProjectAdmin permission set with granular access

Development Workflow

Use feature branches for all development work
Reference docs/onboarding/ for established processes
Leverage realistic sample data for testing scenarios
Follow deployment patterns in scripts/deploy/
All changes validated by GitHub Actions CI/CD

Context Engineering (Patrick's Approach)

Evaluation Criteria for AI-Generated Code

Code Quality: Must pass existing linters and follow Salesforce best practices
Test Coverage: All new Apex classes require >85% test coverage
Bulkification: All SOQL/DML operations must handle collections
Documentation: Include meaningful comments and method descriptions
Security: Follow principle of least privilege, validate inputs

Tools and Standards

Linting: Use Salesforce Code Analyzer for static analysis
Testing: Leverage TestDataFactory patterns for consistent test data
Deployment: Follow scripts/deploy/ patterns for environment promotion
Documentation: Maintain both human (docs/) and AI (memory-bank/) documentation

AI Agent Behavior Guidelines

Before Writing Code: Always explore existing patterns in force-app/
During Planning: Reference memory-bank/ for enterprise patterns
For Testing: Use established TestDataFactory methods
For Deployment: Follow existing script patterns in scripts/deploy/

Execution Framework (Galen's Approach)

Core Methodology: Explore → Plan → Execute

1. EXPLORE Phase

Goal: Build comprehensive context before any code changes
Commands: Use Claude Code to deeply understand the codebase
Context Building: Spend tokens to analyze architecture, patterns, existing code
Key Principle: Never jump straight to execution - always explore first

2. PLAN Phase

Risk Assessment:
- Small tasks: Proceed directly to execution
- Medium tasks: Break into testable, PR-sized chunks
- Large tasks: Create detailed plan with multiple validation rounds
Planning Guidelines:
- Write function names with 1-3 sentence descriptions
- Define test cases in 5-10 words
- Think architecturally, not just at code level

3. EXECUTE Phase

Code Generation: Focus on clean, new code rather than complex edits
Testing: Follow established TestDataFactory patterns
Commits: Use atomic commits with descriptive messages
Validation: Ensure all changes pass existing quality standards

Advanced Context Management

Context Preservation: Use established patterns from memory-bank/
Multiple Sessions: Reference .cursor/rules/ for consistency across sessions
Documentation Updates: Keep both human and AI documentation synchronized

Quality Gates

All code must compile and pass existing tests
Follow conventional commit message format
Maintain consistency with existing architectural patterns
Update relevant documentation (both docs/ and memory-bank/)

AgenticForce Orchestrator

Enterprise Salesforce development template optimized for AI-native workflows with Claude, Cursor, and Cline integration.

Project Context & Purpose

This project serves as a comprehensive template for modern, AI-assisted Salesforce development, demonstrating enterprise-grade practices while maintaining beginner accessibility.

Key Differentiators:

Dual-audience documentation (human vs AI-specific context)
Pre-configured AI development environment with .cursor/ and memory-bank/
Realistic sample data and comprehensive test coverage
Automated CI/CD validation via GitHub Actions

AI Assistant Guidelines

Essential Context Files (Always Reference These)

Enterprise Patterns:
```
memory-bank/enterprise-patterns.md
```
Integration Guidelines:
```
memory-bank/integration-patterns.md
```
Project Overview:
```
memory-bank/project-overview.md
```
AI Rules:
```
.cursor/rules/core.mdc
```
Memory Bank Index:
```
.cursor/rules/memory-bank.mdc
```

When Working on This Project:

Always reference memory-bank/ files for established patterns
Follow .cursor/rules/ for project-specific guidelines
Use existing sample data patterns in data/ folder
Maintain test coverage standards (current: 25KB+ test classes)

Coding Standards:

Follow Salesforce naming conventions and bulkification patterns
Implement comprehensive test coverage (target: >85%)
Use descriptive commit messages following conventional commits
Reference existing architectural patterns before creating new ones

Architecture Overview

Project Structure

├── .cursor/rules/           # AI coding guidelines and context
├── memory-bank/            # Persistent AI knowledge base  
├── force-app/main/default/ # Salesforce metadata
├── docs/onboarding/        # Human-facing documentation
├── data/                   # Sample CSV data for testing
├── scripts/deploy/         # Deployment automation
└── .github/workflows/      # CI/CD validation (now working!)

Salesforce Components

Custom Objects: Project__c with validation rules and automation
Apex Classes: HelloWorld (13KB), TestDataFactory, LoadDataTestExample (6KB)
LWC Components: projectCard for modern, responsive UI
Automation: Flag_Overdue_Projects flow for business logic
Security: ProjectAdmin permission set with granular access

Development Workflow

Use feature branches for all development work
Reference docs/onboarding/ for established processes
Leverage realistic sample data for testing scenarios
Follow deployment patterns in scripts/deploy/
All changes validated by GitHub Actions CI/CD

Context Engineering (Patrick's Approach)

Evaluation Criteria for AI-Generated Code

Code Quality: Must pass existing linters and follow Salesforce best practices
Test Coverage: All new Apex classes require >85% test coverage
Bulkification: All SOQL/DML operations must handle collections
Documentation: Include meaningful comments and method descriptions
Security: Follow principle of least privilege, validate inputs

Tools and Standards

Linting: Use Salesforce Code Analyzer for static analysis
Testing: Leverage TestDataFactory patterns for consistent test data
Deployment: Follow scripts/deploy/ patterns for environment promotion
Documentation: Maintain both human (docs/) and AI (memory-bank/) documentation

AI Agent Behavior Guidelines

Before Writing Code: Always explore existing patterns in force-app/
During Planning: Reference memory-bank/ for enterprise patterns
For Testing: Use established TestDataFactory methods
For Deployment: Follow existing script patterns in scripts/deploy/

Goal: Build comprehensive context before any code changes
Commands: Use Claude Code to deeply understand the codebase
Context Building: Spend tokens to analyze architecture, patterns, existing code
Key Principle: Never jump straight to execution - always explore first

2. PLAN Phase

Risk Assessment:
- Small tasks: Proceed directly to execution
- Medium tasks: Break into testable, PR-sized chunks
- Large tasks: Create detailed plan with multiple validation rounds
Planning Guidelines:
- Write function names with 1-3 sentence descriptions
- Define test cases in 5-10 words
- Think architecturally, not just at code level

3. EXECUTE Phase

Code Generation: Focus on clean, new code rather than complex edits
Testing: Follow established TestDataFactory patterns
Commits: Use atomic commits with descriptive messages
Validation: Ensure all changes pass existing quality standards

Advanced Context Management

Context Preservation: Use established patterns from memory-bank/
Multiple Sessions: Reference .cursor/rules/ for consistency across sessions
Documentation Updates: Keep both human and AI documentation synchronized

Quality Gates

All code must compile and pass existing tests
Follow conventional commit message format
Maintain consistency with existing architectural patterns
Update relevant documentation (both docs/ and memory-bank/)

AgenticForce Orchestrator

Related Skills

Markdown Converter

Nano Banana Pro

1password