vardoger

Connects to your local conversation history from Cursor, Claude Code, Windsurf, Cline, and other AI coding assistants to extract behavioral patterns and generate personalized system prompt additions. Runs entirely on your machine with no external calls. Exposes commands to analyze conversation archives, produce tailored prompts based on your communication style and tech stack, and write them directly to each platform's native rules files. The core flow reads JSONL transcripts, feeds them to your chosen LLM for analysis, and outputs markdown injections that make the assistant progressively better suited to how you code. Useful when you want your AI assistant to remember your preferences without manually maintaining a growing system prompt.