memory-retrieve

Progressive memory loading uses an index-driven filter to scan memory and load only relevant decisions, avoiding large costly memory dumps and reducing latency. It retrieves targeted files on demand during context-building.

Progressive disclosure is a three-level memory loading workflow for durable memory: an index scan, targeted file loading, and optional cross-domain consistency checks. You need it for complex reasoning or multi-task projects requiring historical context.

Cross-domain context retrieval applies optional consistency checks after loading targeted durable-memory files, ensuring contextually relevant decisions are gathered for quick, accurate multi-domain reasoning without full memory dumps.

Yes, memory-retrieve is designed for multi-task projects requiring decision tracking. It loads only relevant DEC-{ID}.md files and related materials on demand, preserving durable memory while providing historical context for accurate reasoning.

The best way to avoid large memory dumps is activating memory-retrieve before context-building. It applies an index-driven filter to load only minimal relevant decisions, reducing latency and preserving durable memory during complex reasoning.

Loading full memory history increases latency because it processes unnecessary data instead of filtering for relevant decisions. Progressive disclosure solves this by scanning an index first, then loading only targeted durable-memory files for the task.