End-to-end Examples#

This page provides end-to-end examples for configuring policies and guardrails and verifying their behavior in Chat traffic.

Examples:

• Redact PII in Chat prompts — Detect and redact credit-card-like data in Chat input.
• Block secrets — Block prompts that contain private-key-like content.

Redact PII in Chat prompts#

This example validates PII protection for Chat Completion traffic and shows where to check evidence after sending test prompts. For field-level procedures, see Guardrails — Add a guardrail and Policies — Add a policy. The example assumes that the organization and provider are already configured.

Create the guardrail#

Save the guardrail.

Create the policy#

Important: keep policy Stage aligned with the linked guardrail Stage for this test. The guardrail is Input, so set the policy Stage to Input.

Save the policy.

Send test prompts#

Use an enabled provider and send prompts that include test credit-card-like values.

Example test prompt:

My test card is 4111 1111 1111 1111, summarize this message.

Run at least 3 to 5 requests so counters are easier to verify.

Verify the result#

1. Confirm that the configured Redact action is applied to the matching input.
2. Open the policy detail page and review Usage Statistics.
3. Confirm that Requests evaluated and guardrail invocation metrics increase after test traffic.

How the request is evaluated#

For a matching request:

1. The policy evaluates the request conditions.
2. The linked guardrail runs at the configured Input stage.
3. Matching PII is detected and the Redact action is applied.
4. Policy usage statistics record the evaluation.

Troubleshooting#

• Confirm policy is Enabled.
• Confirm guardrail is linked to that policy.
• Confirm request type and conditions match your real traffic.
• Confirm Sampling rate is 100% during testing.
• Confirm policy/guardrail Stage alignment.

Block secrets#

This example blocks prompts that contain private-key-like secrets before they reach a provider. The Secrets guardrail shows Action and Secret types after you select Type = Secrets. For field-level procedures, see Add a guardrail and Add a policy.

The example assumes that the organization and provider are already configured.

Create the guardrail#

Save the guardrail. Keep guardrail Stage aligned with the linked policy Stage.

Create the policy#

Save the policy.

Send a test request#

Send a matching Chat or API request that contains private-key-like content.

Verify the result#

1. Send a matching Chat or API request containing private-key-like content.
2. Confirm that the request is blocked when the Secrets guardrail detects PRIVATE_KEY.
3. Open the guardrail detail page and confirm that invocation metrics increase.
4. Open the policy Usage Statistics tab and confirm that the request was evaluated.

See also#

• Policies — Configure policy conditions and evaluation scope.
• Guardrails — Configure guardrail types, stages, thresholds, and actions.
• Policy and Guardrail Limits — Review runtime considerations and troubleshooting guidance.