How to Ask Questions Efficiently

An easy-to-answer question, not necessarily long, but it should let others know what you are doing, what actually happened, and how to see the same problem.

First determine whether it is a question, a reproducible issue, or a feature request, then prepare the corresponding materials.

Choose the appropriate entry first

Your situation	Recommended entry
Not sure whether it's a configuration issue or a product issue	【Settings】→【About Us】→【Feedback】→【Submit with Agent】
Confirmed to be a reproducible bug	GitHub Bug Report
Want to add a new capability	GitHub Feature Request
Discuss usage or solutions	GitHub Discussion / Questions
Only need to look up operation steps	Current documentation and Cherry Assistant

Cherry Assistant can read the current installation package information, and with your consent collect the necessary diagnostics. It will first show a redacted preview, then submit feedback or generate a diagnostic package; you don't need to learn how to read logs first.

What the issue should include

1

1. State the goal in one sentence

For example: “I want the Agent to send the daily report to a Feishu group on weekday mornings.”

2

2. State the actual result and the expected result

Actual result: “The task ran successfully, but Feishu did not receive the message.” Expected result: “After the run completes, the specified group receives one daily report message.”

3

3. Provide the shortest reproduction steps

Start from the entry where the problem can be reproduced stably, and use 3–6 steps to clearly describe the click path, selected content, and where the error appears.

4

4. Add environment details

Specify the operating system, the version shown in 【Settings】→【About Us】, and the relevant service provider or channel type. Do not disclose API Keys or account credentials.

5

5. Add screenshots, screen recordings, or logs

Screenshots should include the error and the page where it occurred; logs should only cover the relevant time period and be redacted first. Issues with model output should include the input, model, and expected format.

Templates you can use directly

Title: Agent scheduled task ran successfully, but the Feishu channel did not receive the result

Goal: Send the operations daily report to the specified Feishu group at 09:00 on weekdays.
Actual result: The run history shows success, but Feishu has no message.
Expected result: A daily report message is received in the group.

Reproduction steps:
1. Open 【Settings】→【Scheduled Tasks】.
2. Select the daily report task and click 【Run】.
3. Wait for the task to complete.
4. Check the Feishu group; there is no new message.

Environment: macOS / the version shown in Cherry Studio 【About Us】 / Feishu channel.
Already tried: re-enabled the channel; sent a test message to the bot in Feishu.
Attachment: screenshot of the channel status with credentials redacted, corresponding run records.

What to remove from screenshots and logs

Before submitting, delete API Keys, Authorization, Cookies, bot Tokens, App Secrets, email addresses, full local paths, chat content, and business data. If a secret is exposed, revoke it immediately from the service provider or platform.

Small habits that speed up troubleshooting

• Reproduce only one problem at a time;
• Use the minimal configuration that can trigger the problem;
• Explain whether the problem occurs consistently;
• Do not just write 'cannot use'; include the original error text;
• Do not upload the entire log if it is unrelated to the problem;
• If you find the same issue, add the environment and new clues instead of creating a duplicate.

Can I let Cherry Assistant submit it for me directly?

Yes. After entering Cherry Assistant through 【Feedback】, explain the issue and explicitly ask it to submit. It will request authorization for diagnostic data, show a redacted preview, and then handle it according to the method you choose. It will only go to a GitHub Issue if you explicitly ask for GitHub.