part-xvi-ai-specific

What problem does it solve?

This Skill defines comprehensive principles, rules, and protocols to govern the behavior, decision-making, and self-correction mechanisms of AI agents, ensuring accuracy, honesty, and reliability in operations.

Core Features & Use Cases

• Anti-Hallucination Protocol: Prevents fabrications and speculations, ensures outputs are factually accurate.
• Verification Hierarchy: Offers a structured approach to verify information, prioritizing confidence levels.
• Self-Correction Mandate: Requires AI agents to proactively review and correct their own work, triggering reviews upon completion or when external cues indicate issues.
• Context Window Management: Ensures efficient use of context and tokens, with clear budgeting and prioritization guidelines.
• Uncertainty Communication Protocol: Standardizes how uncertainty is expressed and handled, aiding in clear communication.
• Mistake Catalog and Response Matrix: Identifies common AI mistakes and outlines response protocols for various scenarios.

Quick Start

Activate the skill unit 'part-xvi-ai-specific' to review and apply the principles of anti-hallucination and self-correction to your AI models and workflows.