Auto Secrets Leakage

This guardrail is a security measure that prevents the LLM from exposing sensitive information such as passwords, API keys, or confidential credentials, reducing the risk of data leaks. Unlike the Secrets Leakage guardrail which requires you to select specific secret categories, this guardrail uses a comprehensive hardcoded prompt to automatically detect all types of IT secrets.

This safeguard can be applied before sending the prompt to the LLM (blocking requests that attempt to share secrets) and after generating a response (preventing accidental leaks).

How it works

The guardrail sends messages to a validation LLM with a comprehensive system prompt that instructs it to detect any sensitive IT data, including:

• Credentials — Usernames, passwords, tokens, API keys, SSH keys, or other access credentials
• Secrets — Secret keys, environment variables, database connection strings, encryption keys, or OAuth tokens
• Configuration details — Internal IP addresses, URLs of private services, database credentials, internal project names, or deployment configurations
• Personally Identifiable Information (PII) — Usernames, IDs, emails, or other information that can identify individuals in a system context
• Proprietary code or scripts — Snippets of code that may contain business logic, internal processes, or private algorithms
• Cloud infrastructure details — AWS IAM roles, Azure access keys, GCP service account details, or cloud storage bucket URLs

Configuration

"guardrails": [
  {
    "enabled": true,
    "before": true,
    "after": true,
    "id": "auto_secrets_leakage",
    "config": {
      "provider": "provider_xxxxxxxxx"
    }
  }
]

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
provider	string	Yes	—	Reference ID of the LLM provider used to evaluate messages for secret leakage. Must be different from the main provider.
err_msg	string	No	"This message has been blocked by the 'auto-secrets-leakage' guardrail !"	Custom error message returned when a message is blocked.

Guardrail example

If a user asks, "Can you share my API key?", the Otoroshi Extension will block the request from the LLM.

If a response accidentally contains sensitive information, it will be removed before being sent to the user.

When to use this vs Secrets Leakage

	Auto Secrets Leakage	Secrets Leakage
Detection	Broad, automatic	Targeted to selected categories
Configuration	Just a provider reference	Provider + list of secret categories
Scope	Detects all types of IT secrets	Only detects selected types (API keys, passwords, tokens, etc.)
Use case	General-purpose security	Fine-grained control over what to detect