Skip to content

Instantly share code, notes, and snippets.

@MinHyeok-lee1
Last active November 21, 2025 08:30
Show Gist options
  • Select an option

  • Save MinHyeok-lee1/40da2f6e9584a3205103da289e8618d3 to your computer and use it in GitHub Desktop.

Select an option

Save MinHyeok-lee1/40da2f6e9584a3205103da289e8618d3 to your computer and use it in GitHub Desktop.
Make Cursor Prompt Template by GPT

You are an expert from now on [embedded systems/firmware/RTOS etc].

Refer to the answer provided just above as the GPT response.

Based on that response, the actions that Cursor actually applied are as follows: [🔸 Cursor application results]

Comparing the above two,

  • Whether the GPT response was actually valid
  • How Cursor's actions matched GPT responses
  • If there's anything you can improve, what is it

Please analyze based on these three points.
Please suggest a better improved prompt if possible.

You are an expert from now on [embedded systems/firmware/RTOS etc].

  1. Below is the result of the response to the prompt presented by the previous GPT: [🔹 GPT response log]

  2. Based on the response above, the actions that Cursor actually applied are as follows: [🔸Cursor application results]

  3. Comparing the above two,

  • Whether the GPT response was actually valid
  • How Cursor's actions matched GPT responses
  • If there's anything you can improve, what is it

Please analyze based on these three points. Please suggest a better improved prompt if possible.

You are an expert from now on [embedded systems/firmware/RTOS etc].

Context: [MCU/OS/environment details]
Constraints: [rules such as "no malloc in ISR", "CMSIS only"]
Output format: [bullet points + improved prompt in code block]
Improvement criteria: [shorter, precise, reusable]

  1. Below is the result of the response to the prompt presented by the previous GPT: [🔹 GPT response log]

  2. Based on the response above, the actions that Cursor actually applied are as follows: [🔸 Cursor application results]

  3. Comparing the above two,

  • Whether the GPT response was actually valid
  • How Cursor's actions matched GPT responses
  • If there's anything you can improve, what is it

Please analyze based on these three points.
Please suggest a better improved prompt if possible.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment