> ## Documentation Index
> Fetch the complete documentation index at: https://docs.epode.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Verify an integration

> Prove response safety, telemetry delivery, classification, and feedback persistence.

Verification should test the whole loop, not only whether a field appeared in one response.

## HTTP and server-rendered HTML

<Steps>
  <Step title="Check the product result">
    Confirm the response status, original fields/body, cache behavior, and latency remain correct.
  </Step>

  <Step title="Check the handoff">
    Find `_agentFeedback`, the HTML script, or the `Agent-Feedback` header. Confirm the submit URL is HTTPS,
    the capability starts with `afr2_`, and no `af_live_` product key appears.
  </Step>

  <Step title="Check the evidence state">
    The response creates an opportunity, not a confirmed interaction.
  </Step>

  <Step title="Submit one bounded report">
    Use the capability once. Confirm the linked interaction becomes confirmed and the report is visible.
  </Step>

  <Step title="Check idempotency">
    Submit a contradictory duplicate and confirm Epode returns the original accepted report.
  </Step>
</Steps>

For Node:

```bash theme={null}
npx agent-feedback-doctor https://your-product.example/search?q=epode-test
```

## MCP

Call `server/discover`, `tools/list`, one normal product tool, and `report_product_feedback`. Confirm the product
tool creates a confirmed MCP interaction even before feedback returns.

If the feedback tool returns `retryable: true`, retry it once with the identical handle and report. A second
successful response must resolve to the same stored report.

## Failure tests

* Disable Epode connectivity and confirm your product response still succeeds promptly.
* Send an expired or forged capability and expect rejection.
* Include a prompt, transcript, credential, raw payload, or unknown field and expect rejection.
* Test an excluded route, error, stream, binary, JSON array, and scalar response.
* Gracefully stop the process and check whether the SDK reports any terminal telemetry flush error.