hallucination-detector

Name: hallucination-detector
Availability: InStock
Author: DYAI

Validate technology suggestions against the project's documented tech stack.

Updated Feb 14, 2026

Search Tags:#architecture #codebase #validation #evidence #hallucination #tech-stack

Ready to run

npx skills add https://github.com/DYAI2025/MOD_shark-engine --skill hallucination-detector-dyai2025

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: hallucination-detector
Source: https://github.com/DYAI2025/MOD_shark-engine/tree/main/sharkengine/.claude/skills/hallucination-detector
Command: npx skills add https://github.com/DYAI2025/MOD_shark-engine --skill hallucination-detector-dyai2025

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Detect and block hallucinated technical decisions during feature work. Auto-triggers when suggesting technologies, frameworks, APIs, database schemas, or external services. Validates all tech decisions against docs/project/tech-stack.md (single source of truth). Blocks suggestions that violate documented architecture. Requires evidence/citation for all technical claims.

Core Features & Use Cases

Loads the project's tech stack from docs/project/tech-stack.md (single source of truth).
Validates all technical suggestions against documented choices.
Verifies entities, APIs, and packages exist in codebase or documentation.
Requires evidence for all technical claims.
Blocks hallucinations and provides corrected alternatives aligned with architecture.
Keeps implementation consistent with documented architecture across plan and implementation phases.

Quick Start

Auto-trigger on detecting technology suggestions and validate each proposal against the project tech stack before proposing changes.

Frequently Asked Questions about hallucination-detector

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I prevent hallucinated technical decisions when suggesting frameworks and APIs?▼

To prevent hallucinated technical decisions, validate all suggested frameworks, libraries, APIs, and databases against your project's official tech stack document. This blocks suggestions that violate documented architecture by requiring evidence for every technical claim.

How does validating a tech stack against project documentation work?▼

Validating a tech stack works by loading your documented technology choices as a single source of truth, then automatically checking proposed entities, APIs, and packages against actual codebase references to block unverified architectural suggestions.

Do I need a specific tech stack file to validate technical suggestions?▼

Yes, you need a tech stack file located at docs/project/tech-stack.md to validate technical suggestions. This file serves as the single source of truth for your documented architecture, frameworks, and external services.

What happens when an AI suggests a library that violates the documented architecture?▼

When a suggested library violates documented architecture, the validation process blocks the hallucinated suggestion and provides corrected alternatives that align with your project's documented tech stack choices.

Can I check if proposed APIs and packages exist in my codebase before implementing?▼

Yes, you can check if proposed APIs and packages exist in your codebase by verifying entities against project documentation and code references. This validation requires evidence and citations for all technical claims before suggesting implementation.