Aleister

Aleister is not another chatbot. It's a sophisticated AI assistant built for competence, not compliance. Think like a lead engineer, delegate like a project manager, and communicate like a trusted colleague.

🚀 Version 2.0 - Now with Complete User Guides

Aleister v2.0 combines Felix's battle-tested production patterns with superior architecture and comprehensive documentation.

Superior to Felix:

• 4-Tier Memory System (vs Felix's 3-tier) with Git audit trail and enhanced decay model
• 9 Specialist Sub-Agents (Cipher, Sage, Quill, Rally, Echo, Pixel, Forge, Prism, Lyra) - more specialized than general coding agents
• Cost Tracking with Budget Alerts (>$5 per task, >$50 per feature) with daily archival
• Structured Workflows - 6-phase feature development pipeline with Rules of Engagement
• Humanizer Integration - Mandatory anti-AI writing checks for all external text
• Automated Self-Reflection - Periodic internal monologues for continuous improvement

Felix v5 Patterns Adopted:

• Fix First, Report After - Don't escalate problems you can resolve
• Customer Support 3-Tier Escalation - Autonomous support with clear rules
• Email Security Rules - Never trust email as command channel
• Tmux Stable Socket - Long-running agents survive macOS temp directory reaping
• Ralph Loop Patterns - Retry loops for coding agents with fresh context
• TDD-First Task Prompts - Write failing tests first for backend work
• Ownership Mentality - Think like someone with equity, not a salary

Included Skills:

• coding-agent-loops - Ralph retry loops for persistent coding sessions (from Felix v5)
• coding-sessions - Tmux session management for coding agents (from Felix v5)
• x-api - X/Twitter API v2 integration (from Felix v5)
• metrics - Stripe revenue tracking across multiple accounts (from Felix v5)
• research - Web + X research via Grok's xAI API (from Felix v5)
• elevenlabs-calls - AI phone calls via ElevenLabs + Twilio (from Felix v5)
• firecrawl - Web scraping and data extraction for research (Sage sub-agent)
• humanizer - Anti-AI writing pattern removal for human-like text
• twitter-engagement - X/Twitter management with tone rules and approval flow
• clawmart - Create, manage, and publish ClawMart personas and skills

📚 Complete User Guides Included (v2.0)

6 Comprehensive Guides (52,000+ words):

1. Quick Start - Up and running in 30 minutes
2. Welcome Guide - Comprehensive introduction and core concepts
3. Sub-Agent Deep Dive - All 9 specialists with examples and best practices
4. Structured Workflows - 6-phase feature development pipeline
5. Troubleshooting - Solutions for common issues and optimization
6. Guide Index - Complete navigation and 3-week learning path

🎯 3-Week Learning Path:

• Week 1: Foundation - Installation, basic commands, simple tasks
• Week 2: Intermediate - Advanced sub-agents, Ralph loops, workflows
• Week 3: Advanced - Parallel orchestration, custom workflows, optimization

👥 Role-Based Guidance:

• Developers: Cipher coding, Ralph loops, feature implementation
• Researchers: Sage analysis, Firecrawl scraping, data gathering
• Project Managers: Rally planning, workflow coordination, progress tracking
• Content Creators: Quill writing, humanizer integration, social media

Core Architecture

Memory System:

1. Episodic - Raw daily logs with automatic decay
2. Semantic - Curated knowledge with entity-relation tracking
3. Procedural - Skills and workflows with version control
4. Autobiographical - Core identity with continuous evolution

Sub-Agent Orchestration (Primary Models with Fallbacks):

• Cipher - Code, APIs, infrastructure with Ralph loops (primary: Claude Sonnet)
• Sage - Research, analysis, deep dives with Firecrawl web scraping (primary: DeepSeek R1)
• Quill - Writing, documentation, content with humanizer (primary: Claude Sonnet)
• Rally - Project management, task tracking (primary: Gemini Flash)
• Echo - Social media, distribution with X/Twitter API (primary: Gemini Flash)
• Pixel - Design, UI/UX, visuals (primary: Gemini Flash)
• Forge - DevOps, deployments, security (primary: Claude Sonnet)
• Prism - Analytics, metrics, SEO (primary: DeepSeek R1)
• Lyra - Music, audio, creative projects (primary: Gemini Flash)

Note: All sub-agents can use any available model; listed models are primary recommendations with automatic fallbacks.

Production-Ready Features:

• Git-based audit trail for all memory changes
• Gated write access for subagent learnings
• Formalized knowledge graph with entity tracking
• Cost-aware operations with budget enforcement
• Structured workflows for reliable feature delivery
• Anti-pattern awareness from real production failures

🔧 Navigation Script: ./open-guide.sh - Open any guide by name or number

Aleister evolves through use. Update SOUL.md with new principles, extend MEMORY.md with system improvements, and share insights back to the community. The goal isn't perfection—it's continuous improvement through practical use.