FSAI - Claude Code Plugins

Claude Code plugins for code quality and development workflows.

Installation

# Add the marketplace
/plugin marketplace add franchiseai/claude-code-plugins

# Install plugins
/plugin install code-cleanup@fsai
/plugin install dead-code@fsai
/plugin install worktree@fsai
/plugin install architecture-refactor@fsai
/plugin install entity-graph@fsai
/plugin install ux-flows@fsai
/plugin install fsai-workflow-actions@fsai

Plugins

code-cleanup

Remove AI-generated code patterns ("slop") from a branch.

Usage: Just ask Claude Code:

• "Remove the slop from this branch"
• "Clean up AI code"
• "Make this code look human-written"

Claude will diff against master, identify AI patterns, and fix them.

---

dead-code

Remove dead code from a branch.

Usage: Just ask Claude Code:

• "Remove the dead code from this branch"

Claude will diff against master, identify dead code, and remove it.

---

worktree

Manage git worktrees for parallel development.

Usage: Just ask Claude Code:

• "Create a worktree for feature-xyz"
• "List my worktrees"
• "Remove the fsai2 worktree"

Claude will manage worktrees and remind you to install dependencies.

---

architecture-refactor

Analyze and refactor backend code to follow the entity/service separation pattern.

The Pattern:

Controller (HTTP only)
    ↓
Service (orchestration, enforcement, side effects)
    ↓
Entity (business rules as predicates, data access)
    ↓
Database

Usage: Just ask Claude Code:

• "Refactor deals to entity pattern"
• "Analyze architecture for leads"
• "Extract business rules from application service"
• "Where are the rules for deals?"
• "Create entity for franchisee"
• "Separate concerns for locations"

What it does:

1. Analyzes the current service/controller for a domain concept
2. Identifies scattered business rules, data access, and side effects
3. Proposes an entity interface with predicates and data methods
4. Refactors the service to be an orchestrator that:
   • Calls entity predicates (canConvert(), isEntityDeal())
   • Enforces rules (throws if predicate returns false)
   • Coordinates multiple entities in transactions
   • Triggers side effects after success

Key Principles:

Layer	Does	Does Not
Entity	Define rules as predicates (canX() → boolean), CRUD, queries	Throw errors, call other entities
Service	Orchestrate, enforce rules (throw), business logic, transactions	Define rules, write SQL
Controller	Parse request, auth, call service, format response	Business logic, call entities

Example Transformation:

Before (rules scattered in service):

async convertDeal(dealId) {
  const deal = await this.getDealOverview(dealId);
  if (deal.convertedAt) throw new ValidationError('Already converted');
  if (!deal.applicationId) throw new ValidationError('No application');
  // ... 200 lines of mixed logic
}

After (rules in entity, service orchestrates):

async convertDeal(dealId) {
  const deal = await dealEntity.getOverview(dealId);

  const { allowed, reason } = dealEntity.canConvert(deal);
  if (!allowed) throw new ValidationError(reason);

  // Orchestrate...
  await dealNotifications.onConverted(deal);
}

---

entity-graph

Design data models as Mermaid ER diagrams with implementation-ready Drizzle schema code.

Usage: Just ask Claude Code:

• "Design the data model for campaigns"
• "I need new tables for user notifications"
• "Schema design for the scheduling feature"

Claude will read your existing schema conventions, ask clarifying questions, present a Mermaid ER diagram, and write an implementation plan with exact Drizzle code.

---

ux-flows

Turn meeting notes and technical specs into clean UX flow docs for design teams.

Usage: Just ask Claude Code:

• "Clean up these flows for the design team"
• "Write UX flows from these meeting notes"
• "Design handoff doc for the onboarding feature"

Claude will strip implementation details and layout prescriptions, keep edge cases and states, and output a ~1 page doc in user-facing language.

---

fsai-workflow-actions

FSAI workflow actions and content generation tools. Includes a content generator skill that helps produce valid JSON import files for the FSAI applicant portal and email sequences.

Usage: Just ask Claude Code:

• "Generate portal JSON for this brand"
• "Create email sequences"
• "FSAI content generator"

Or invoke directly with /content-generator.

Prerequisites: Export a brand metadata JSON from the FSAI admin panel and place it in metadata/ in your workspace. The skill reads asset IDs, form names, and brand context from this file.

What it does:

1. Reads brand metadata (assets, forms, brand info)
2. Guides content generation following the portal and sequence JSON schemas
3. Validates output with a bundled validate.ts script
4. Catches broken asset references, invalid form names, and schema errors before import

---

Contributing

To add a new plugin:

1. Create a folder in plugins/ with your plugin name
2. Add plugin.json with metadata
3. Add skill.md with the instruction set
4. Update marketplace.json to include your plugin