Store and retrieve cross-session lessons in a SQLite FTS5 memory database.
npx skills add https://github.com/z1439527767/claude-config --skill memory-lora
Please help me install this Agent Skill. Skill: memory-lora Source: https://github.com/z1439527767/claude-config/tree/main/skills/imported/memory-lora Command: npx skills add https://github.com/z1439527767/claude-config --skill memory-lora
This Skill solves the problem of losing important lessons, fixes, decisions, and reusable patterns between AI sessions by providing a structured cross-session memory system.
Use the memory-lora skill to store and recall important lessons from my current development workflow.
High-intent search queries and answers about installing and using this skill.
Cross-session memory is preserved by storing bugs, corrections, and lessons in a SQLite database with FTS5 search. This enables AI workflows to retrieve reusable task outcomes and maintain project continuity across sessions.
FTS5 search retrieves relevant past lessons by scoring memories through keyword, temporal, and entity-based strategies. It tracks usefulness over time to ensure accurate error recovery and pattern discovery during task execution.
Yes, SQLite with FTS5 is required to manage durable memory records and perform intelligent recall. It provides the necessary full-text search and scoring mechanisms to store and retrieve cross-session corrections effectively.
The best way to build a knowledge retrieval system for recurring errors is using a structured memory database that applies scoring mechanisms. This approach preserves insights so future sessions can automatically apply learned solutions.
Yes, cross-session learning tracks pattern discovery by storing task outcomes and applying temporal scoring strategies. This monitors memory usefulness over time, ensuring that discovered patterns remain accessible for future AI assistant workflows.
SQLite limits AI memory management to local FTS5 full-text search and scoring strategies rather than distributed vector databases. Memory recall depends entirely on keyword, temporal, and entity-based matching without semantic vector search.
malue-aiStore and search personal knowledge snippets in a SQLite FTS5 database.
SherifEldeebCreate structured incident records, timelines, evidence, and lessons learned.
ArogyaReddyCombine vector similarity and keyword matching with RRF, Linear, Cross-encoder, or Cascade fusion.
DammianMillerManage a 4-layer memory system using SQLite and Qdrant for AI coding agents.
