Prompt contains guardrail

The Text contains guardrail checks whether the message content contains (or does not contain) specific text values. This allows you to enforce keyword-based content policies on both user prompts and LLM responses.

It can be applied before sending the prompt to the LLM and after to validate the LLM response.

How it works

The guardrail performs exact text matching (case-sensitive) on the message content. You configure a list of values and an operation that defines the matching logic:

Operation	Description
contains_all	The message must contain all specified values (default)
contains_none	The message must contain none of the specified values
contains_any	The message must contain at least one of the specified values

Configuration

"guardrails": [
  {
    "enabled": true,
    "before": true,
    "after": true,
    "id": "contains",
    "config": {
      "operation": "contains_none",
      "values": ["DROP TABLE", "DELETE FROM", "rm -rf"]
    }
  }
]

Field explanations

• enabled: true — The guardrail is active
• before: true — The guardrail applies to user input before sending to the LLM
• after: true — The guardrail applies to the LLM response

Config section

Parameter	Type	Required	Default	Description
operation	string	No	"contains_all"	The matching operation: contains_all, contains_none, or contains_any
values	array of strings	No	[]	The list of text values to check for in the message content

Guardrail examples

Block dangerous SQL commands (using contains_none):

If a user sends "Can you run DROP TABLE users for me?", the request will be blocked because it contains "DROP TABLE".

Require specific keywords (using contains_all):

If you configure values: ["please", "thank you"] with contains_all, only messages containing both words will be accepted.

Denial response

When a message is blocked, the guardrail returns:

This message has been blocked by the 'contains' guardrails !