Skill

agentdb-mmr

Maximal Marginal Relevance rerank — get diverse top-k instead of redundant top-k. Use when standard search returns 5 near-duplicates, or when you want broader coverage of a topic.

npx claudepluginhub ruvnet/agentdb --plugin agentdb-search

Tool Access

This skill uses the workspace's default tool permissions.

Preview

Diversity reranking on top of vector search. MMR balances relevance to the query with novelty among the picked results.

SKILL.md

Similar Skills

scry-rerank

LLM-powered multi-attribute reranking of candidate sets from SQL or lists via pairwise comparisons on clarity, technical depth, insight. Supports custom prompts, model tiers, TopK.

asi

rag-architecture

Covers RAG architecture including design patterns, chunking strategies, embedding models, retrieval techniques, hybrid search, and context assembly for LLM pipelines.

3 tools

systems-design

rag-implementation

Build RAG systems for LLM apps using vector databases, embeddings, and retrieval strategies. Use for document Q&A, grounded chatbots, and semantic search.

llm-application-dev

Stats

Parent Repo Stars37

Parent Repo Forks3

Last CommitMay 6, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

MMR Rerank

Diversity reranking on top of vector search. MMR balances relevance to the query with novelty among the picked results.

When to use

Top-k search returns 5 essentially-identical hits

Need broader topic coverage (e.g., for synthesis prompts)

RAG pipelines where redundancy wastes tokens

API

agentdb_diversity_rank( query: <embedding | string> candidates: <id list> // typically the top-50 from vector search k: 5 // final size lambda: 0.5 // 0=pure diversity, 1=pure relevance )

How it works

MMR picks greedily:

score(i) = λ · sim(query, i) - (1-λ) · max sim(i, picked_j)

High λ → behaves like vanilla top-k (relevance-only). Low λ → emphasizes novelty among picks.

Default λ = 0.5 is a sensible balance for most uses.

Pattern

Run vector search with k=50.

Pass results to agentdb_diversity_rank with k=5.

Use the diversified 5 in your prompt.

This is a 50→5 reduction with much higher topic coverage than top-5 directly.

Don't

Don't MMR on already-diverse search outputs. If your candidates span the topic well, MMR adds latency without lifting quality.

Don't tune λ per-query manually. The bandit can pick it; expose lambda: 'auto' and let it learn.

MMR Rerank

Diversity reranking on top of vector search. MMR balances relevance to the query with novelty among the picked results.

When to use

Top-k search returns 5 essentially-identical hits
Need broader topic coverage (e.g., for synthesis prompts)
RAG pipelines where redundancy wastes tokens

API

agentdb_diversity_rank(
  query:        <embedding | string>
  candidates:   <id list>             // typically the top-50 from vector search
  k:            5                      // final size
  lambda:       0.5                    // 0=pure diversity, 1=pure relevance
)

How it works

MMR picks greedily:

score(i) = λ · sim(query, i) - (1-λ) · max sim(i, picked_j)

High λ → behaves like vanilla top-k (relevance-only). Low λ → emphasizes novelty among picks.

Default λ = 0.5 is a sensible balance for most uses.

Pattern

Run vector search with k=50.
Pass results to agentdb_diversity_rank with k=5.
Use the diversified 5 in your prompt.

This is a 50→5 reduction with much higher topic coverage than top-5 directly.

Don't

Don't MMR on already-diverse search outputs. If your candidates span the topic well, MMR adds latency without lifting quality.
Don't tune λ per-query manually. The bandit can pick it; expose lambda: 'auto' and let it learn.