smart-explore

Memory-guided codebase exploration uses a vectorized memory index to query past architecture notes and recent decisions, prioritizing targeted file reads over blind scanning to quickly build an accurate project mental model. It surfaces prior context to avoid duplicating work.

Developer onboarding via memory-guided exploration retrieves timeline data of recent edits and past decisions, selectively reads relevant files, generates a project overview summarizing purpose and current state, and saves new insights back to memory.

Resuming paused work leverages stored session memory to retrieve prior decisions and recent file history, enabling targeted context-building that skips redundant file scanning and immediately surfaces known issues and next steps.

Project exploration requires a vectorized memory index for querying past architecture notes, alongside tools for timeline retrieval of recent edits, file reading capabilities, and functionality to persist new findings back into memory.

Memory-guided codebase auditing handles context-building by querying vectorized memory for past decisions, performing targeted file reads, and capturing insights by producing a concise project overview that is persisted back into memory for future sessions.