What are the different intent classifications used?

Queries are classified into four categories: Explore (broad discovery), Answer (precision-focused), Locate (exact matching), and Time-boxed (date-filtered).

Can I see why a specific search method was chosen?

Yes, every operation returns a detailed explainability payload highlighting the tier used, layers attempted, execution time, and confidence scores.

What is the primary purpose of the retrieval-policy skill?

It serves as the decision-making layer that determines the most effective search strategy for agent memory queries based on intent and available data layers.

How does it handle failed search attempts?

The skill uses predefined fallback chains, automatically transitioning from advanced layers like Hybrid search to simpler methods like BM25 or Agentic search if primary methods fail.

Intelligent Retrieval Policy

Name: Intelligent Retrieval Policy
Author: SpillwaveSolutions

bySpillwaveSolutions

•

数据科学与机器学习

Orchestrates complex memory search operations by classifying query intent and routing requests through optimal retrieval layers.

The Retrieval Policy skill acts as the decision-making engine for agentic memory systems, ensuring Claude uses the most effective search strategy for any given request. By analyzing query intent—ranging from broad exploration to specific time-boxed lookups—the skill automatically selects between topic-based, hybrid, vector, or keyword search layers. It ensures high reliability through automated fallback chains and provides full transparency with detailed explainability payloads that reveal exactly how and why a search result was retrieved.

主要功能

014 GitHub stars

02Intelligent intent classification (Explore, Answer, Locate, Time-boxed)

03Comprehensive explainability payloads for search transparency

04Multi-tier retrieval routing across Hybrid, Vector, and BM25 layers

05Customizable stop conditions for depth, nodes, and timeouts

06Automated fallback chains for high-reliability memory access

使用场景

01Debugging agent reasoning by analyzing the explainability reports of memory retrievals.

02Optimizing search performance by dynamically selecting the most efficient retrieval layer.

03Ensuring search reliability in complex agent environments through automated fallback mechanisms.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add spillwavesolutions/agent-memory retrieval-policy

For use in Claude.ai and ChatGPT

Download Skill