3.7 KiB
3.7 KiB
Date/Time
Friday, October 24, 2025 (Timezone: UTC+00:00 - Please adjust to local system time)
Change Tracking/Revision Table
| Date | Version | Description | Author |
|---|---|---|---|
| 2025-10-24 | 1.0.0 | Initial documentation for AI tool preferences | AIOS-Public System |
Changelog
| Date | Version | Description |
|---|---|---|
| 2025-10-24 | 1.0.0 | Initial creation of AI tool usage documentation |
AI Tool Preferences and Usage Patterns
Overview
This document captures the current AI tool preferences and usage patterns for the AI home directory environment. It provides context for any AI agent working within this environment.
Primary Tools and Usage
1. Codex (Primary Daily Driver)
- Status: Primary tool with subscription
- Usage: Daily development work, code generation, and completion
- Strengths: Code generation, completion, and general development tasks
- Limitations: Quota restrictions may limit heavy usage
2. Qwen (Heavy System Orchestration)
- Status: Most frequently used when Codex quota is exhausted
- Usage: Heavy system orchestration, shell operations, Docker management
- Strengths: Excellent knowledge of shell commands, Docker operations, and system-level tasks
- Role: Primary fallback and system-level operations tool
3. Gemini (Audit Specialist)
- Status: Primary audit and analysis tool
- Usage: Code audits, analysis, and review tasks
- Strengths: Analytical capabilities, detailed reviews, comparison operations
- Role: Quality assurance and audit functions
4. Claude (Occasional Use)
- Status: Used occasionally for specific tasks
- Usage: Specific reasoning tasks when needed
- Strengths: Advanced reasoning, document analysis, creative tasks
- Role: Specialized tasks when other tools aren't optimal
Usage Patterns
Daily Workflow
- Morning: Start with Codex (if quota available) for primary development tasks
- Throughout Day: Switch to Qwen when Codex quota is reached or for system operations
- Review/Audit: Use Gemini for detailed analysis and code reviews
- Special Tasks: Use Claude for specific reasoning requirements
Tool Selection Guidelines
- Code Generation/Completion: Codex (primary), Qwen (when Codex unavailable)
- System Orchestration: Qwen (primary), Codex (secondary)
- Audits and Reviews: Gemini (primary), others (as needed)
- Advanced Reasoning: Claude (primary for complex reasoning), others (as needed)
Context for AI Agents
When working in this environment, AI agents should be aware that:
- Commands and operations may be executed by different AI tools depending on availability
- Qwen is particularly valuable for shell and Docker operations
- Codex is the preferred tool for code generation when available
- Gemini should be considered for audit and analysis tasks
- The tool selection may change based on subscription quotas and availability
Performance Considerations
Tool Characteristics
- Codex: Excellent for code generation but limited by subscription quota
- Qwen: Reliable for system operations, good for heavy tasks, no quota limits
- Gemini: Best for analytical tasks and detailed reviews
- Claude: Good for complex reasoning but used selectively
Optimizing Tool Usage
- Leverage Qwen for operations that require extensive system knowledge
- Use Codex for code-related tasks when available
- Employ Gemini for validation and quality checks
- Reserve Claude for specialized reasoning tasks