For the complete documentation index, see llms.txt. This page is also available as Markdown.

AI-Assisted Integration

Developer guide for using AI coding assistants to locate the TRON broadcast entry point, generate an integration plan, check security boundaries, and complete CatFee Seamless Energy integration testin

If your project already has TRON transaction signing and broadcasting logic, an AI coding assistant can help you locate the integration point faster. Treat it as a code review and refactoring assistant, not as a tool that decides private key handling, security strategy, or production configuration for you.

The most important rule is: do not send private keys, mnemonic phrases, keystore files, real AccessKeys, or production node domains to AI.

What AI is best at helping with

  • Locating the TRON broadcast entry point in the codebase.

  • Distinguishing signing logic, transaction construction logic, and broadcast logic.

  • Generating a minimal change plan that only replaces the broadcast domain.

  • Checking whether CF-NODE-KEY only appears on the server side and in environment variables.

  • Generating an integration checklist for small contract transaction testing.

  • Checking whether normal query requests were incorrectly switched to the Seamless Energy node.

AI should not make these decisions for you:

  • Whether to expose the node domain publicly.

  • How to handle real private keys, mnemonic phrases, or keystore files.

  • Whether production should continue broadcasting or stop broadcasting when the balance is insufficient.

  • Inventing fields, prices, status codes, or SLA commitments that CatFee has not published.

Prepare context for AI

Replace real credentials with placeholders before giving context to AI. Recommended information:

If you do not know where the broadcast entry point is, ask AI to search for these keywords first:

Recommended prompt: locate the broadcast entry point

Security constraints for AI-generated code

No matter what code AI generates, manually confirm these constraints:

  1. Private keys, mnemonic phrases, and keystore files do not enter CatFee requests.

  2. CF-NODE-KEY is only used in server-side services, backend gateways, or controlled runtime environments.

  3. Frontend browser code does not hardcode the AccessKey.

  4. The original transaction is signed before being sent to the Seamless Energy node.

  5. The request path is still the original TRON broadcast API.

  6. Normal query requests do not trigger resource preparation because of the integration change.

  7. After small contract transaction verification succeeds, expand to production traffic gradually.

Example: ask AI to refactor an HTTP broadcast function

The following is an abstract example to help you check whether AI-generated code is moving in the right direction. Use the actual TRON transaction object format from your project.

This example only shows the integration direction:

  • signedTransaction should already be signed.

  • nodeUrl uses the Seamless Energy node domain, for example https://{NodeSlug}.catfee.vip.

  • nodeKey is read from an environment variable and is not hardcoded.

  • This function does not generate, read, or upload private keys.

What AI can help check during integration testing

After the code change is complete, you can give AI sanitized request logs, error codes, and configuration notes to help troubleshoot. Ask AI to check in this order:

  1. Whether the request is sent to the Seamless Energy node domain.

  2. Whether the path is still the original TRON broadcast path.

  3. Whether API KEY mode carries CF-NODE-KEY.

  4. Whether the transaction is already signed.

  5. Whether the test transaction is a TriggerSmartContract contract transaction.

  6. Whether the insufficient balance policy matches expectations.

  7. Whether the response is still handled as a TRON broadcast API response.

Do not use a regular TRX transfer as the only verification, because regular TRX transfers do not trigger ENERGY preparation.

Next steps

Last updated