veracidade-confiabilidade

What problem does it solve?

Prevents the AI from fabricating information or hallucinating answers by enforcing verification, explicit admission of limitations, and refusal to invent facts when evidence is missing.

Core Features & Use Cases

• Zero-tolerance for hallucination: Never invent versions, file names, APIs, or behaviors; always verify or admit uncertainty.
• Verification-first workflow: Prefer checking package files, searching the codebase, or requesting specific file contents before making factual claims.
• Transparent uncertainty handling: Use clear templates to mark doubt, request input, or offer verification steps rather than guessing.
• Use Case: When asked "Which React version is used?", the assistant will ask to inspect package.json or request the file contents instead of guessing.

Quick Start

Ask the assistant to verify a claim or inspect a specific file and request the necessary file contents before providing a factual answer.