If you're dealing with fragmented memory systems across your Claude workflows, this skill consolidates everything into a unified AgentDB backend with HNSW vector search. The pitch is 150x to 12,500x faster search performance compared to whatever you're running now, presumably SQLite or markdown-based storage. It handles the migration automatically while keeping your existing integrations working. The speed gains sound aggressive, but HNSW is legit for vector similarity search if you're doing heavy retrieval work. Worth testing if memory lookups are a bottleneck in your agent setup, though you'll want to verify those benchmarks against your actual data patterns.
npx -y skills add ruvnet/ruflo --skill "V3 Memory Unification" --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills