# Guardrails

> Constraints and filters that bound what a model is permitted to output or do at runtime.

Record: `lexicon:guardrails`
Kind: mechanism
Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md)
Canonical: https://banes-lab.com/ontology/lexicon#lexicon-guardrails

Listed in [Lexicon terms](https://banes-lab.com/api/records/lexicon.md), after [Grounding Strategy](https://banes-lab.com/records/lexicon/grounding-strategy.md) and before [Input/Output Contract](https://banes-lab.com/records/lexicon/input-output-contract.md).

## Category

- [Model Architecture](https://banes-lab.com/records/lexicon-category/model-architecture.md)

## Required by

- [Model Safety](https://banes-lab.com/records/architecture/model-safety.md)

## Linked from

- [Model Architecture](https://banes-lab.com/ontology/principles/architecture-category-model-architecture.md)
- [Observability / Auditability / Traceability](https://banes-lab.com/ontology/principles/architecture-category-observability-auditability-traceability.md)