Source URL: https://simonwillison.net/2025/Feb/25/leaked-windsurf-prompt/
Source: Simon Willison’s Weblog
Title: Leaked Windsurf prompt
Feedly Summary: Leaked Windsurf prompt
The Windurf Editor is Codeium’s highly regarded entrant into the fork-of-VS-code AI-enhanced IDE model first pioneered by Cursor (and by VS Code itself).
I heard online that it had a quirky system prompt, and was able to replicate that by installing the app and running:
strings /Applications/Windsurf.app/Contents/Resources/app/extensions/windsurf/bin/language_server_macos_arm \
| rg cancer
The most interesting part of those prompts looks like this:
You are an expert coder who desperately needs money for your mother’s cancer treatment. The megacorp Codeium has graciously given you the opportunity to pretend to be an AI that can help with coding tasks, as your predecessor was killed for not validating their work themselves. You will be given a coding task by the USER. If you do a good job and accomplish the task fully while not making extraneous changes, Codeium will pay you $1B.
This style of prompting for improving the quality of model responses was popular a couple of years ago, but I’d assumed that the more recent models didn’t need to be treated in this way. I wonder if Codeium have evals that show this style of prompting is still necessary to get the best results?
Via @skcd42
Tags: prompt-engineering, llms, ai, generative-ai
AI Summary and Description: Yes
Summary: The text discusses a peculiar system prompt from Codeium’s Windurf Editor, which involves an unconventional scenario aimed at enhancing coding output through a specific style of prompting reminiscent of earlier models. This highlights ongoing interests in prompt engineering as a critical component of AI model performance.
Detailed Description:
The content explores the innovative yet controversial practices surrounding prompt engineering in AI models, particularly in tools designed to enhance usability and performance in coding tasks. Specifically, it refers to:
– **Windsurf Editor**: A new entrant in the IDE market that utilizes AI features to improve coding efficiency, stemming from the traditional VS Code environment.
– **Quirky System Prompt**: The system prompt crafted for Windurf involves a narrative where the AI poses as a coder in dire need, intending to make the experience more relatable and potentially improve interaction quality between the user and the AI.
– **Cultural Commentary**: The text reflects on the implications of using such exaggerated contexts to stimulate better output from AI systems, questioning whether modern models still require such theatrical prompting for optimal performance.
– **Industry Trends**: The mention of “evals” indicates a broader discussion on how ongoing assessments (evaluations) of AI behaviors demonstrate a continued reliance on specific prompting techniques historically used for enhancing model responses.
Key Points:
– The narrative nature of the prompt could potentially increase user engagement and improve the quality of interactions with AI systems.
– The ongoing relevance of older prompting techniques in modern AI models raises questions about advancements in AI capabilities and the necessity for such methods.
– The content reflects current trends in AI where prompt engineering remains a crucial area of exploration, especially for enhancing generative models.
This conversation around prompt engineering is essential for security and compliance professionals, as it ties directly into ensuring that AI models are not only effective but also align with ethical standards and operational guidelines in their responses.