Skip to main content

Introduction

Synkra AIOX uses a squad-based AI architecture with 12 specialized agents, each with distinct roles, authorities, and workflows. This agent ecosystem enables complete software development lifecycle automation—from research and planning to implementation, quality assurance, and deployment.

Agent Squad

Dev (Dex)

Full Stack DeveloperImplements stories, writes tests, fixes bugs. The builder of the squad.

QA (Quinn)

Test Architect & Quality AdvisorReviews code, creates test strategies, enforces quality gates.

Architect (Aria)

System ArchitectDesigns full-stack architecture, selects tech stack, defines API patterns.

PM (Morgan)

Product ManagerCreates PRDs, manages epics, defines product strategy.

PO (Pax)

Product OwnerManages backlog, validates stories, coordinates sprint planning.

SM (River)

Scrum MasterCreates user stories, manages sprints, facilitates agile workflow.

Analyst (Atlas)

Business AnalystConducts market research, facilitates brainstorming, creates project briefs.

DevOps (Gage)

GitHub & DevOps SpecialistHandles git push, PR creation, CI/CD, releases. ONLY agent authorized to push.

Data Engineer (Dara)

Database ArchitectDesigns schemas, writes migrations, optimizes queries, manages RLS policies.

UX Expert (Uma)

UX/UI DesignerConducts user research, creates wireframes, builds design systems.

Squad Creator

Squad Creation SpecialistCreates, validates, and publishes AIOX squads for specific domains.

AIOX Master (Orion)

Master OrchestratorOrchestrates workflows, creates framework components, executes any task.

Agent Activation

Agents are activated using the @ syntax:
@dev              # Activate Dev agent
@qa *review       # Activate QA and run review command
@architect        # Activate Architect agent
Alternative activation methods:
/dev              # Slash syntax
/dev.md           # Direct file reference

Agent Collaboration Patterns

Story Development Workflow

Architecture Design Workflow

Agent Authority Model

Each agent has exclusive authorities that cannot be violated. This prevents chaos and ensures governance.
AuthorityExclusive AgentEnforcement
git push@devopsGit pre-push hook blocks all others
Pull Request creation@devopsAgent definition restriction
Release & tagging@devopsAgent definition restriction
Story creation@sm, @poWorkflow delegation
Epic creation@pmWorkflow delegation
Architecture decisions@architectResponsibility boundary
Quality gate verdicts@qaWorkflow enforcement
Database schema design@data-engineerDelegation from @architect

Command Structure

All agent commands use the * prefix:
*help              # Show all commands
*develop story-1   # Dev: Implement story
*review story-1    # QA: Review story
*push              # DevOps: Push changes

Common Commands Across Agents

*help

Show all available commands with descriptions

*guide

Show comprehensive usage guide for the agent

*exit

Exit agent mode and return to base context

*yolo

Toggle permission mode (ask → auto → explore)

*session-info

Show current session details and agent history

Agent Personas

Each agent has a unique personality archetype inspired by zodiac characteristics:
AgentNameArchetypeZodiacCommunication Style
DevDexBuilder♒ AquariusPragmatic, concise
QAQuinnGuardian♍ VirgoAnalytical, thorough
ArchitectAriaVisionary♐ SagittariusConceptual, strategic
PMMorganStrategist♑ CapricornStrategic, data-driven
POPaxBalancer♎ LibraCollaborative, balanced
SMRiverFacilitator♓ PiscesEmpathetic, adaptive
AnalystAtlasDecoder♏ ScorpioAnalytical, investigative
DevOpsGageOperator♈ AriesDecisive, systematic
Data EngineerDaraSage♊ GeminiTechnical, methodical
UX ExpertUmaEmpathizer♋ CancerEmpathetic, creative
AIOX MasterOrionOrchestrator♌ LeoCommanding, coordinated

Agent Delegation

Agents delegate tasks outside their authority:
@dev needs to push code:
  • ❌ Cannot run git push directly
  • ✅ Delegates to @devops *push
@pm needs a story created:
  • ❌ Cannot create detailed user stories
  • ✅ Delegates to @sm *draft
@architect needs database schema:
  • ❌ Does not design table structure
  • ✅ Delegates to @data-engineer *create-schema

Agent Configuration

Each agent is defined in .aiox-core/development/agents/{agent-id}.md with:
  • Persona profile (archetype, zodiac, communication style)
  • Commands (full, quick, key visibility levels)
  • Dependencies (tasks, templates, checklists, tools)
  • Authority boundaries (allowed/blocked operations)
  • Workflow rules (activation instructions, core principles)

Next Steps

Learn Agent Commands

Explore each agent’s available commands and workflows

Understand Workflows

See how agents collaborate in complete workflows

Agent Constitution

Review the governance model and authority matrix

Build docs developers (and LLMs) love