I would like to request the addition of a
Persistent Context
feature in Windsurf, similar to the functionality available in the Codeium Extension.
**Codeium Extension: Codeium Persistent Context**
This feature would allow users to define custom rules and guidelines for the AI to follow, enhancing the relevance and accuracy of AI-generated code suggestions.
Features:
  • Custom Chat Instructions
    : A short prompt guideline like “Respond in Kotlin and assume I have little familiarity with it” to orient the model towards a certain type of response. This would help in tailoring the AI's responses to the user's specific needs and expertise level.
Here are my actual custom instructions:
>
"Prioritize efficient, error-free code, ensure clarity, readability, and adherence to style guides. Always think step-by-step, creating a plan before coding. Prioritize optimal, efficient solutions. Include brief explanations/comments and handle potential edge cases. Seek clarity on inputs/outputs and ask follow-ups for vague requests."
  • Pinned Contexts
    : Items from the codebase such as files, directories, and code snippets that you would like the model to explicitly take into account. Here there can be multiple pinned contexts, e.g. docuemtation or the "Rules for AI" (see feature request Rules for AI). This ensures that the AI has access to the most relevant parts of the project. See also Context Pinning.
  • Active Document(s)
    : A marker for the currently active file(s), which receives special focus. This allows the AI to prioritize the context of the file Cacsade is actively working on, improving the relevance of suggestions.
  • Local Indexes
    : A list of local repositories that the Codeium context engine has indexed. This feature enables the AI to leverage a broader context from the local codebase, enhancing its ability to provide accurate and contextually relevant suggestions. See also Local Indexes.
Feel free to reach out if you have any questions or feedback on this feature request. https://www.watunga.ai