Skip to main content

User Quick Start Guide

User Quick Start Guide

Get from zero to productive in 10 minutes.

CODITECT is not just for software development. With 130+ specialized agents and 750+ workflows, it handles business strategy, market research, competitive intelligence, education, content creation, HR, legal, finance, and virtually any knowledge work.

Prerequisites

  • Git 2.25+
  • Python 3.10+
  • Claude Code (Anthropic CLI)

Step 0: Environment Setup (First Time Only)

# Create virtual environment
python3 -m venv .venv

# Activate it
source .venv/bin/activate # macOS/Linux
# .venv Scripts activate # Windows

# Install dependencies
pip install -r requirements.txt

# Verify
which python # Should show .venv/bin/python

Step 1: Guided Onboarding (5 minutes)

Start the interactive onboarding consultant:

/onboard

The coditect-onboarding agent will:

  • Assess your experience level
  • Verify your environment (Git, Python, Claude Code)
  • Set up your shell with the AI Command Router
  • Create your first project
  • Teach you the Task Tool Pattern

Onboarding options:

/onboard           # Full guided experience
/onboard --quick # Skip assessments, use defaults
/onboard --resume # Resume from last checkpoint
/onboard --status # Check your progress

Step 2: Create Your First Project (2 minutes)

# Standard project creation
/new-project "Build a task management API"

# Guided mode with mentor support (recommended for new users)
/new-project --guided "Build a task management API"

The --guided flag activates the first-project-companion agent with:

  • Session check-ins with progress context
  • Milestone celebrations
  • Help when you're stuck
  • Gradual fade-out as you gain confidence

What you get:

  • Complete Git repository
  • Production folder structure (60+ points)
  • project-plan.md with milestones
  • tasklist.md with checkbox tracking
  • CI/CD workflows (GitHub Actions)
  • .coditect symlink for AI intelligence

Step 3: Set Up AI Command Router (1 minute)

Add to your shell config (~/.zshrc or ~/.bashrc):

alias cr='claude "/suggest-agent"'
alias cri='claude --interactive "/suggest-agent"'

Reload:

source ~/.zshrc  # or source ~/.bashrc

Now describe what you want in plain English:

cr "I want to add user authentication"
# → Suggests: /implement or /security-hardening

cr "analyze my database schema"
# → Suggests: database-architect agent

cr "help me debug this error"
# → Suggests: debugger agent

Step 4: Invoke Your First Agent (2 minutes)

The Task Tool Pattern:

Task(
subagent_type="agent-name",
prompt="Detailed description of what you want"
)

Try this:

# Business analysis
Task(
subagent_type="codi-business-analyst",
prompt="Create Product-Market Fit analysis for my SaaS platform"
)

# Technical architecture
Task(
subagent_type="cloud-architect",
prompt="Design scalable infrastructure for my application"
)

# Multi-agent orchestration
Task(
subagent_type="orchestrator",
prompt="Coordinate complete feature development from requirements to deployment"
)

Key rules:

  1. subagent_type = the agent name directly (e.g., "orchestrator", "Explore", "debugger")
  2. prompt = detailed description with specific requirements
  3. Include context, expected outputs, and file paths for best results

Daily Workflow

Start of Session

# Quick context check
git log --oneline -5

# Query recent activity
/cxq --recent 50

# Recall specific topic
/cxq --recall "what you're working on"

# Check recent decisions
/cxq --decisions --limit 10

During Work

# Search for prior work
/cxq "authentication implementation"

# Find code patterns
/cxq --patterns --language python

# Find error solutions
/cxq --errors "TypeError"

# Check decisions made
/cxq --decisions --decision-type api

End of Session

# 1. Export session transcript (creates readable backup)
/export

# 2. Process and index (captures to searchable database)
/cx

# Combined: export + process + index in one step
/cxs

Why this matters: /export creates a human-readable session transcript. /cx extracts unique messages and indexes them for search. /cxs does both - always run before ending work.

Learn the Platform with /cxq

Use /cxq to explore the knowledge base and discover CODITECT features.

Find Agents and Commands

# Discover agents by task
/cxq "agent for database"
/cxq "agent for security"
/cxq "agent for deployment"

# Learn command usage
/cxq "git-sync examples"
/cxq "new-project options"
/cxq "how to deploy"

Find Code Examples

# Implementation patterns
/cxq "REST API example"
/cxq "authentication pattern"
/cxq "database connection"

# By language
/cxq --patterns --language python
/cxq --patterns --language typescript
/cxq --patterns --language rust

Troubleshoot Problems

# Find error solutions
/cxq --errors "ModuleNotFoundError"
/cxq --errors "connection refused"
/cxq "how to fix permission denied"

# Prior debugging sessions
/cxq "resolved bug"
/cxq "test fix"

Explore Framework Features

# Understand concepts
/cxq "what is orchestrator"
/cxq "skill vs command difference"
/cxq "component activation explained"

# Find workflows
/cxq "workflow for code review"
/cxq "multi-agent coordination"
/cxq "CI/CD pipeline setup"

Query by Category

# Decisions and architecture
/cxq --decisions # All decisions
/cxq --decisions --decision-type api
/cxq --decisions --decision-type security

# Time-based queries
/cxq --today # Today's activity
/cxq --recent 100 # Last 100 messages
/cxq --recall "project name" # Topic recall

Tip: Your knowledge base grows with every session. Search often to leverage prior work.

Essential Commands Cheat Sheet

Session Management

/cxs                             # Save and index session context
/session-index --verbose # Scan all sessions for discovery
/process-jsonl-sessions --batch # Process large session files

Project Organization

/new-project "description"                # Create new project
/new-project --guided "description" # With mentor support
/organize-production --mode analyze # Validate structure
/organize-production --mode validate # Get readiness score (0-100)

Git Workflows

/git-sync --target all --mode full  # Full repository sync
/git-sync --target all --dry-run # Preview changes
/git-sync --mode analyze # Analysis only

Component Management

# Activate a component
python3 scripts/update-component-activation.py activate agent NAME --reason "purpose"

# Check status
python3 scripts/update-component-activation.py status agent NAME

# List activated components
python3 scripts/update-component-activation.py list --activated-only

Testing

# Run test suite
python3 scripts/test-suite.py

# Run specific tests
pytest tests/test_onboarding_system.py -v

# Run with coverage
pytest --cov tests/

The 7 Component Types

CODITECT has 7 types of components - here's the 30-second summary:

ComponentHow to UseExample
Commands/command-name/new-project, /git-sync
AgentsTask(subagent_type=...)orchestrator, debugger
SkillsAutomatic (background)git-workflow-automation
Scriptspython3 scripts/name.pytest-suite.py
ToolsClaude uses automaticallyRead, Write, Grep, Bash
WorkflowsMulti-step processesFeature development
CookbookReady recipes15 common patterns

Full guide: See CODITECT-WELCOME-ABOARD.md for detailed component explanations.

Recipe 1: Start a New API Project

Task(
subagent_type="orchestrator",
prompt="Create a complete project plan for a REST API with user authentication, CRUD operations, and PostgreSQL database"
)

Recipe 2: Debug an Error

Task(
subagent_type="debugger",
prompt="Investigate this error: [paste error message]. Check logs, trace the issue, and suggest fixes."
)

Recipe 3: Review Code Quality

Task(
subagent_type="code-reviewer",
prompt="Review the changes in [file/directory] for security, performance, and best practices"
)

Recipe 4: Deploy to Production

Task(
subagent_type="devops-engineer",
prompt="Create a CI/CD pipeline with GitHub Actions, Docker containerization, and Kubernetes deployment"
)

Recipe 5: Market Analysis (Business)

Task(
subagent_type="business-intelligence-analyst",
prompt="Create comprehensive market analysis for [your market] including TAM/SAM/SOM sizing, competitive landscape, and pricing analysis"
)

Recipe 6: Competitive Intelligence (Research)

Task(
subagent_type="competitive-market-analyst",
prompt="Research top 5 competitors to [your product]. Analyze their pricing, features, target audience, and market positioning. Create comparison matrix."
)

Recipe 7: Course Curriculum (Education)

Task(
subagent_type="ai-curriculum-specialist",
prompt="Design a 5-module beginner curriculum for [topic]. Include learning objectives, hands-on exercises, assessments, and estimated completion times."
)

Recipe 8: Investment Readiness (Finance)

Task(
subagent_type="venture-capital-business-analyst",
prompt="Prepare Series A readiness assessment for [startup]. Include valuation analysis, unit economics review, and investor pitch recommendations."
)

See all 30 recipes: CODITECT Cookbook

Training Pathways

Quick Path (30 minutes)

  1. Complete /onboard guided setup
  2. Create first project with /new-project --guided
  3. Invoke 3 agents using Task Tool Pattern
  4. Read CODITECT-COOKBOOK.md

Standard Path (2-3 hours)

  1. Complete Quick Path above
  2. Read WHAT-IS-CODITECT-CORE.md
  3. Practice all recipes from Cookbook
  4. Complete Module 1-2 from training system

Certification Path (4-6 hours)

  1. Complete Standard Path above
  2. Read USER-TRAINING-PATHWAYS.md
  3. Complete all 5 modules with assessments
  4. Build sample project end-to-end
  5. Pass certification exam (80%+ required)

Key Concepts

Agents (130+ specialists)

Specialized AI experts for every domain - not just software development:

Software Development

AgentUse For
orchestratorMulti-agent coordination, complex workflows
ExploreFast codebase exploration
debuggerBug investigation and fixes
code-reviewerCode quality reviews
security-specialistSecurity audits
devops-engineerCI/CD, deployment
database-architectDatabase design

Business & Strategy

AgentUse For
business-intelligence-analystMarket sizing, financial modeling, unit economics
competitive-market-analystCompetitive intelligence, market positioning
venture-capital-business-analystInvestment readiness, valuations, pitch prep
content-marketingMarketing strategy, content planning

Research & Intelligence

AgentUse For
claude-research-agentDeep research synthesis
web-search-researcherCurrent information, web research
biographical-researcherPeople/company research profiles

Education & Content

AgentUse For
ai-curriculum-specialistCourse design, learning paths
assessment-creation-agentQuizzes, evaluations, rubrics
codi-documentation-writerTechnical docs, guides, knowledge bases

Commands (141 available)

Quick actions starting with /:

/new-project    # Create new project
/git-sync # Sync repositories
/cxs # Save session
/deploy # Deploy to production
/analyze # Code analysis

Component Activation

Components require explicit activation before use:

python3 scripts/update-component-activation.py activate agent orchestrator \
--reason "Multi-agent workflow coordination"

Why? Prevents context pollution, faster startup, clear audit trail.

Getting Help

Documentation

Community

Support

  • Email: support@az1.ai (AZ1.AI customers)
  • Commercial support plans available

What's Next?

After completing this quick start:

  1. Memory System - Never lose context

  2. Multi-Agent Orchestration - Coordinate complex workflows

  3. Complete Training - Master all CODITECT capabilities

  4. Workflow Library - 750+ ready workflows across 15 domains

  5. All Recipes - 30 ready-to-use patterns

  6. Full Capabilities Guide - See everything CODITECT can do


Welcome to CODITECT - where AI-powered development becomes effortless.

You've got this!


Last Updated: December 22, 2025 Version: 1.0.0 Framework: CODITECT v1.7.2