hallucination-detector

Hallucination detection validates all technology suggestions against your documented tech stack and existing codebase. It compares recommended frameworks, libraries, APIs, and databases against docs/project/tech-stack.md to block suggestions that don't align with your project's actual architecture and require evidence for every claim.

Yes. The validator loads your tech stack documentation, verifies that suggested frameworks, libraries, and APIs actually exist in your codebase, and blocks recommendations that violate your documented architecture. All suggestions must be backed by citations from your existing technology choices.

Validation compares suggested database schemas and APIs against your project's documented tech stack and existing codebase. Mismatches are caught and corrected with evidence-backed alternatives that reflect your actual project technology.

The detector blocks violations and provides corrected alternatives that align with your tech stack. Every suggestion requires citation from your docs/project/tech-stack.md or existing codebase, preventing unsupported architectural decisions.

Yes. The validator requires an up-to-date docs/project/tech-stack.md file that lists your frameworks, libraries, databases, APIs, and cloud services. This reference document is compared against all technology recommendations during feature work.

Yes. Detection applies to frameworks, libraries, APIs, database schemas, cloud services, and architectural patterns. All technology recommendations—whether during feature development or architectural decisions—are validated against your documented tech stack and existing codebase.