# Hulk

Keep calm: you have a Hulk with you now. Enhanced log filtering, multi-header/params API credentials, AI··Agent usability and security, and offline detection

## NEW

### Enhanced audit log filtering 📊

Ref: #4213

We've upgraded the log search for you! Now, you can quickly filter through your logs. Choose from `statuses` like Failure, Success, or All, and `triggers` such as Manual, Emailhook, Webhook, etc., to narrow down your search effortlessly.

   <div class="docs-screenshot-compact">

![Enhanced audit log filtering](./assets/enhanced-audit-log-filtering.png)

   </div>

   <div class="docs-screenshot-compact">

![Enhanced audit log filtering](./assets/enhanced-audit-log-filtering-2.png)

   </div>

### Support multiple headers and query parameters in API key credential **🛠**

Ref: #4096

We've improved the API key credential type by enabling the addition of multiple key/value pairs in both headers and query parameters.

This enhancement is crucial because different services and APIs require various methods of authentication for secure access. By supporting a broader range of headers and query parameters, our platform can accommodate more complex authentication schemes, ensuring seamless integration with a wider array of third-party services.

   <div class="docs-screenshot-compact">

![Support multiple headers query parameters api](./assets/support-multiple-headers-query-parameters-api.png)

   </div>

### Silent errors: network offline/online detection **🌐**

Ref: #4095

To enhance your experience and minimize the risk of losing work due to connection issues—like those encountered during travel—we've introduced a feature that alerts you when your network goes offline or comes back online. This proactive notification aims to prevent frustration and ensure a smoother workflow, keeping you informed and your work safe.

   <div class="docs-screenshot-compact">

![Silent errors](./assets/silent-errors.png)

   </div>

## CHANGES

### AI··Agent: Restrict sensitive operations **🚫**

Ref: #3962

We've implemented a safety measure for AI··Agents, setting their default permissions to "Read Only" (GET) actions. This precaution is designed to prevent unintended consequences, ensuring that AI operations are safe and controlled. If needed, you can customize this behavior by adjusting the permissions in the “Settings” tab of the AI··Agent.

   <div class="docs-screenshot-compact">

![Ai agent](./assets/ai-agent.png)

   </div>

### AI··Agent: Dry run **🏃**

Ref: #3262

AI··Agent steps have been upgraded to be on par with other workflow steps, meaning they can now undergo dry runs! This enhancement allows you to test and refine AI··Agent steps within your workflows.

   <div class="docs-screenshot-compact">

![Ai agent](./assets/ai-agent-2.png)

   </div>

## FIXES

- Fixed an issue where forwarded emails wouldn't trigger a workflow (Ref: #4142).
- Resolved a problem with extracting data from emails (Ref: #4193).
- Corrected support for custom properties in specific APIs (Ref: #4217).
- Addressed a bug causing HTTP response bodies to copy as a storage handle instead of the actual data (Ref: #4194).
- Fixed a blank page error when assigning credentials to ‘none’ workspace (Ref: #4192).
- Various bug fixes for AI··Agent (Ref: #4108).
- The search bar now remains above playbooks when scrolling (Ref: #4102).
- Fixed issues with copying, pasting, or importing playbooks with nested loops (Ref: #4069).
- When copying large JSON data, it now correctly copies the actual JSON instead of a base64 payload (Ref: #3959).
- Resolved compatibility issues with some AWS services, including problems invoking Lambda and pattern mismatches (Ref: #3171).