Package Exports
This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (@defai.digital/automatosx) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.
Readme
AutomatosX
AI Agent Orchestration for Claude Code
Transform Claude Code into a multi-agent powerhouse with persistent memory, intelligent delegation, and zero-cost knowledge management.
Status: โ Production Ready ยท v5.0.8 ยท October 2025
๐ฏ Built for Claude Code
AutomatosX extends Claude Code with specialized AI agents that remember context, delegate tasks, and collaborate autonomously.
# In Claude Code, simply use /ax
/ax run paris "Design authentication system with JWT"
/ax run sofia "Implement the auth design" # Sofia auto-receives Paris's design from memory
/ax memory search "authentication" # Instant search of all past decisions
The result: Claude Code becomes a learning, coordinated team instead of a stateless assistant.
๐ก Why AutomatosX?
The Problem with Stateless AI
Traditional AI assistants (ChatGPT, vanilla Claude):
- โ No memory between conversations
- โ You repeat context every time
- โ No coordination between tasks
- โ Knowledge disappears after each session
AutomatosX changes this:
- โ Persistent memory: < 1ms search, zero cost, 100% local
- โ Multi-agent delegation: Agents coordinate automatically
- โ Context retention: Never explain the same thing twice
- โ Knowledge accumulation: Your team gets smarter over time
Real-World Impact
Without AutomatosX:
Day 1: You explain architecture to Claude โ Response lost
Day 2: You ask to implement โ You re-explain architecture
Day 3: Different task โ You re-explain everything again
With AutomatosX:
Day 1: Paris designs architecture โ Saved to memory
Day 2: /ax run sofia "implement auth" โ Sofia finds Paris's design automatically
Day 3: /ax run steve "security audit" โ Steve has full context from Day 1-2
Time saved: Hours per week. Quality: Consistent. Cost: $0.
๐ What's New
v5.0.8 (October 2025): Critical Fixes - Timeout & Memory
- CRITICAL FIX: Multi-stage agents now respect
--timeout
flag - CRITICAL FIX: Memory system enforces
maxEntries
andautoCleanup
limits - Timeout support: AbortSignal properly passed to all stage executors
- Memory limits: Automatic cleanup prevents database growth issues
- 100% backward compatible: Drop-in replacement for v5.0.7
v5.0.6: File Operation Tools Enabled v5.0.5: Provider Parameters & Version Management v5.0.4: Memory saving now works automatically v5.0.3: Special character support in memory search v5.0.0: Agent template system for quick agent creation
๐ Full Changelog | ๐ Release Notes
๐ง Core Value: Persistent Memory
AutomatosX remembers everything. Every agent conversation is automatically saved and searchable.
How It Works
// Automatic memory saving
/ax run paris "Design calculator with add/subtract"
โ Task + Response saved to SQLite FTS5
// Automatic memory retrieval
/ax run sofia "Implement the calculator"
โ Memory searches "calculator" automatically
โ Sofia receives: "# Relevant Context from Memory: Paris's design..."
โ Sofia implements WITHOUT you repeating the spec
The Technology
- SQLite FTS5: Built-in full-text search
- < 1ms search: 62x faster than v3.x vector search
- $0 cost: No embedding APIs, no cloud calls
- 100% local: Your data never leaves your machine
- Automatic injection: Relevant context added to every agent
Benefits
โ Cross-day continuity: Pick up where you left off โ Cross-agent knowledge: All agents share the same knowledge base โ Learning from history: Agents avoid past mistakes โ Zero cost scaling: 10,000 entries = ~10MB, still < 1ms search
Learn more: Memory System Guide | Memory Tutorial
๐ค Core Value: Multi-Agent Orchestration
Agents coordinate automatically. Natural language delegation creates complex workflows without manual orchestration.
How It Works
// Product Manager analyzes and delegates
/ax run paris "Build authentication feature"
Paris response:
"I'll design the auth system with JWT + OAuth2.
@sofia Please implement the JWT authentication API based on this design.
@steve Please audit the implementation for security issues."
// AutomatosX automatically:
// 1. Sofia receives full spec, implements code
// 2. Steve receives spec + code, performs audit
// 3. Results aggregated back to Paris
The Technology
- 7 delegation syntaxes:
@mention
,DELEGATE TO
,Please ask
, etc. - Cycle detection: Prevents infinite loops
- Depth limits: Default 2 levels (configurable)
- Session tracking: Who did what, when
- Workspace isolation: No file collisions
Benefits
โ Automatic coordination: No manual task switching โ Parallel execution: Multiple agents work simultaneously โ Transparent workflows: Full delegation chain visible โ Context preservation: Every agent has complete context
Learn more: Multi-Agent Orchestration Guide
๐ญ 15 Specialized Agents, 4 Professional Teams
Every agent optimized for their domain with the best AI provider:
๐ฅ Core Team (OpenAI)
General assistance and code generation
- Alex - Versatile assistant for general tasks
- Sofia - Senior software engineer (clean code, TDD, pragmatic)
- Ryan - Code reviewer (quality, security, performance)
- Danny - Debugger specialist
- Wendy - Technical writer
๐ป Engineering Team (Claude)
Deep reasoning for technical work
- Bob - Backend expert (API design, databases, microservices)
- Frank - Frontend specialist (React, UX, accessibility)
- Oliver - DevOps engineer (infrastructure, CI/CD, cloud)
- Steve - Security expert (threat modeling, security audit)
- Queenie - QA specialist (testing strategies, test automation)
๐ Business Team (Gemini)
Strategic thinking and analysis
- Eric - CEO (business strategy, organizational leadership)
- Tony - CTO (technology strategy, technical leadership)
- Daisy - Data Analyst (data analysis, ML, statistical modeling)
๐จ Design Team (Gemini)
Creative and design work
- Paris - Product Manager (product strategy, user research)
- Debbee - UX/UI Designer (user experience, visual design)
โก Quick Start
Installation
npm install -g @defai.digital/automatosx
In Claude Code
# Initialize (first time only)
/ax init
# Run agents
/ax run paris "Design REST API for users"
/ax run sofia "Implement the API" # Auto-receives Paris's design
/ax run queenie "Write tests for the API" # Auto-receives design + implementation
# Search memory
/ax memory search "API design"
/ax memory list --agent paris
# Manage agents
/ax agent list
/ax agent show sofia
/ax agent create backend --template developer
That's it! Agents now remember everything and coordinate automatically.
๐ Full Installation Guide | Quick Start Tutorial
๐ Documentation
Getting Started
- Quick Start Guide - Get up and running in 5 minutes
- Core Concepts - Understand agents, memory, providers
- Installation Guide - Detailed setup instructions
Core Features
- Agent Communication & Memory - How agents communicate and remember
- Multi-Agent Orchestration - Natural language delegation
- Team Configuration - Team-based agent organization
- Agent Templates - Quick agent creation
Tutorials
- Memory Management - Hands-on memory system guide
- Creating Your First Agent - Build custom agents
Reference
- CLI Commands - Complete command reference
- Agent Directory - All available agents
๐ฌ The Technical Advantage
Feature | Traditional AI Chat | Claude Code | Claude Code + AutomatosX |
---|---|---|---|
Memory | No | No | โ SQLite FTS5 (< 1ms) |
Cost | $20/month | Included | โ $0 (100% local) |
Multi-Agent | No | No | โ 15 specialized agents |
Coordination | Manual | Manual | โ Automatic delegation |
Context Retention | Copy-paste | Session only | โ Persistent (days/weeks) |
Knowledge Sharing | No | No | โ Cross-agent memory |
Privacy | Cloud | Claude servers | โ 100% local data |
Speed | Web UI | Terminal | โ Instant CLI |
๐ผ Real-World Use Cases
๐๏ธ Feature Development
/ax run paris "Design user authentication feature"
# Paris creates spec โ Saved to memory
/ax run sofia "Implement auth based on spec"
# Sofia auto-receives spec โ Implements code
/ax run steve "Security audit the auth implementation"
# Steve auto-receives spec + code โ Performs audit
/ax run wendy "Document the auth system"
# Wendy auto-receives everything โ Creates docs
Result: 4-step workflow, zero context re-explanation, complete audit trail
๐ Bug Investigation
/ax run danny "Debug the payment timeout issue"
# Danny analyzes, saves findings to memory
/ax run sofia "Fix the issue Danny found"
# Sofia reads Danny's analysis โ Implements fix
/ax run queenie "Test the payment fix"
# Queenie knows the bug + fix โ Comprehensive testing
Result: Coordinated debugging with full context preservation
๐ Research & Analysis
/ax run daisy "Analyze user behavior patterns"
# Daisy analyzes data โ Findings in memory
/ax run paris "Design features based on Daisy's analysis"
# Paris reads analysis โ Creates product spec
/ax run eric "Business case for Paris's proposal"
# Eric has analysis + spec โ Strategic evaluation
Result: Data-driven decision making with complete context
๐ฏ Why Teams Choose AutomatosX
For Solo Developers
- Extend Claude Code with persistent memory
- Never repeat context - agents remember everything
- Coordinate complex tasks with multi-agent workflows
- 100% local - your data stays private
For Teams
- Shared knowledge base - export/import memory across team
- Consistent quality - agents learn from past work
- Faster onboarding - new members inherit team knowledge
- Audit trail - complete history of all decisions
For Claude Code Power Users
- Slash command integration -
/ax
for instant access - Terminal-native - no context switching
- CLI-based - scriptable and automatable
- Zero latency - local memory = instant search
๐ ๏ธ Production-Ready
โ 1,149 tests passing (100% pass rate) โ TypeScript strict mode (zero errors) โ 84% test coverage (comprehensive testing) โ 46MB bundle (87% smaller than v3.x) โ < 1ms memory search (62x faster than v3.x)
Performance Metrics
Memory Search: < 1ms (10,000 entries)
Bundle Size: 46MB (down from 340MB in v3.x)
Dependencies: 158 packages (down from 589 in v3.x)
Test Coverage: 84.19% (1,149 tests)
Memory Cost: $0 (no API calls)
Technology Stack
- Runtime: Node.js 20+
- Language: TypeScript 5.3 (strict mode)
- Memory: SQLite + FTS5 (built-in full-text search)
- Testing: Vitest 2.x (1,149 tests)
- Build: tsup/esbuild
- Providers: Claude CLI, Gemini CLI, OpenAI Codex
๐ง Coming Soon
- Enhanced Claude Code integration
- Visual workflow builder
- Advanced memory analytics
- Cross-project knowledge sharing
- Plugin system for custom providers
๐ค Contributing
We welcome contributions! AutomatosX is built in the open.
- Contributing Guide - How to contribute
- Development Setup - Local setup
- Architecture Guide - Understand the codebase
Join the community:
- ๐ Report Issues
- ๐ก Feature Requests
๐ License
AutomatosX is Apache 2.0 licensed.
๐ Links
- ๐ฆ npm: @defai.digital/automatosx
- ๐ GitHub: defai-digital/automatosx
- ๐ Documentation: docs/
- ๐ Releases: GitHub Releases
- ๐ Changelog: CHANGELOG.md
Transform Claude Code into an intelligent, coordinated team with AutomatosX. ๐
Built with โค๏ธ by the AutomatosX team