The file transfer logic in the workflow orchestration page is inconsistent with that of external API calls.

### Self Checks

- [x] I have read the [Contributing Guide](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md) and [Language Policy](https://github.com/langgenius/dify/issues/1542).
- [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general).
- [x] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones.
- [x] I confirm that I am using English to submit this report, otherwise it will be closed.
- [x] 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
- [x] Please do not modify this template :) and fill in all the required fields.

### Dify version

1.16.1

### Cloud or Self Hosted

Self Hosted (Docker)

### Steps to reproduce

1.Create a workflow that takes a URL as input and outputs an image description.
2.Register this tool in the new version of the agent.
3.Register the workflow as a tool
4.Provide a sample URL for testing in new agent.

### ✔️ Expected Behavior

The vision model receives the image in the workflow, processes it, and generates an image description.

### ❌ Actual Behavior

The model cannot access the image and therefore fails to parse it, but it works fine when test-run within the workflow orchestration page.

Reproduce the issue through the workflow orchestration page, then follow the same URL input through the registered workflow and agent tool paths. Compare their file-transfer behavior with external API calls and verify that the vision model receives the image and produces its description when the workflow is used as a tool.

Written by the indexing model from the issue text.