Memory Search
- 661 installs
- 67k repo stars
- Updated August 4, 2026
- ruvnet/ruflo
memory-search is a Ruflo agent skill that runs hybrid semantic retrieval across stored agent memory so developers who need prior decisions, code patterns, and session context can query them with dense, Graph RAG, or MMR-
About
memory-search is a state-of-the-art semantic retrieval skill for the Ruflo agent memory system, exposing hybrid sparse+dense search, Graph RAG multi-hop traversal, MMR diversity reranking, and recency weighting through Claude Flow MCP tools like memory_search, memory_store, memory_retrieve, and agentdb_context-synthesize. Developers reach for memory-search when a coding agent must recall architecture choices, naming conventions, or debugging conclusions from earlier sessions instead of re-scanning the repository. Strategy selection is query-driven: dense for fast single-hop lookups, --hybrid for keyword-heavy questions, and --graph-rag for multi-hop reasoning across linked memory nodes. Namespace scoping keeps team or project memories isolated while unified search spans stored artifacts.
- Semantic memory retrieval for long-running agent sessions
- Reduces repetitive explanations and context loss
- Enables agents to reference their own prior outputs and decisions
- Compatible with Claude Code, Cursor, and other agentic workflows
- Installs as a lightweight MCP server from ruvnet/ruflo
Memory Search by the numbers
- 661 all-time installs (skills.sh)
- +7 installs in the week ending Aug 5, 2026 (Skillselion tracking)
- Ranked #1,480 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/ruvnet/ruflo --skill memory-searchAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 661 |
|---|---|
| repo stars | ★ 67k |
| Last updated | August 4, 2026 |
| Repository | ruvnet/ruflo ↗ |
How do agents search prior session memory semantically?
Let their coding agent retrieve relevant past decisions, code patterns, and context from previous sessions.
Who is it for?
Developers running long-horizon Claude Flow or Ruflo agents who need durable, searchable memory across sessions.
Skip if: Developers who only need one-shot codebase grep or ripgrep search without persistent agent memory storage.
When should I use this skill?
The agent must recall prior architectural decisions, stored patterns, or cross-session context before changing code.
What you get
Ranked memory snippets, linked context chains, and namespace-scoped retrieval results for agent prompts.
- ranked memory snippets
- synthesized context blocks
- namespace-filtered recall results
By the numbers
- Exposes 7 Claude Flow MCP memory and agentdb tools
- Supports 3 retrieval strategies: dense, hybrid, and Graph RAG multi-hop
Files
Memory Search (SOTA)
State-of-the-art semantic search across Ruflo memory with multiple retrieval strategies.
Strategy Selection
Choose based on query type:
- Default (dense): fast single-hop semantic match
- --hybrid: sparse + dense with RRF fusion (20-49% better for keyword+semantic queries)
- --graph-rag: multi-hop knowledge retrieval (30-60% better for reasoning queries)
Steps
1. Parse query and flags — extract search text and strategy flags from arguments 2. Select retrieval strategy:
Dense search (default):
npx @claude-flow/cli@latest memory search --query "QUERY" --namespace NAMESPACE --limit 10Or via MCP: mcp__claude-flow__memory_search({ query: "QUERY", namespace: "NAMESPACE", limit: 10 })
Hybrid search (when --hybrid or query has specific keywords):
npx ruvector search "QUERY" --hybrid --limit 10Graph RAG (when --graph-rag or multi-hop reasoning needed):
npx ruvector search "QUERY" --graph-rag --limit 10Smart retrieval (when --smart or complex recall needed):
npx @claude-flow/cli@latest memory search --query "QUERY" --smart --limit 10Or via MCP: mcp__claude-flow__memory_search({ query: "QUERY", smart: true, limit: 10 })
Applies 5-phase pipeline: query expansion, RRF fusion, recency boost, MMR diversity, session round-robin. Best for: multi-session recall, temporal queries, diverse result sets.
Unified cross-namespace: mcp__claude-flow__memory_search_unified({ query: "QUERY", limit: 10 })
3. Apply MMR reranking — for diverse results, filter near-duplicates (cosine > 0.92) while maximizing relevance 4. Apply recency weighting — boost recent entries with exponential decay (0.95/day) 5. Synthesize context (for complex queries): mcp__claude-flow__agentdb_context-synthesize({ query: "QUERY", sources: ["patterns", "tasks", "solutions"] }) 6. Present results — ranked by composite score (relevance diversity recency), with source namespace attribution
Namespace Guide
| Namespace | Best For |
|---|---|
patterns | "How did we handle X?" |
tasks | "What was the context for Y?" |
solutions | "How did we fix Z?" |
feedback | "What did the user prefer?" |
security | "Known vulnerabilities in..." |
| (omit) | Search all namespaces |
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10How it compares
Pick memory-search over plain codebase search when decisions and rationale live in agent memory rather than source files.
FAQ
What retrieval modes does memory-search support?
memory-search supports dense default search, hybrid sparse+dense via --hybrid, Graph RAG multi-hop via --graph-rag, plus MMR diversity reranking and recency weighting. Choose dense for speed, hybrid for keyword queries, and graph-rag for linked context chains.
Which MCP tools does memory-search call?
memory-search integrates Claude Flow MCP tools including memory_search, memory_store, memory_list, memory_retrieve, memory_search_unified, agentdb_pattern-search, and agentdb_context-synthesize for unified retrieval and synthesis.