What happens if the primary search layer fails?

The skill follows a predefined fallback chain, automatically attempting the next best search method (e.g., moving from Vector to BM25) to ensure the agent doesn't return empty results.

What is a retrieval tier?

A retrieval tier represents the available search capabilities of your system, ranging from a 'Full' cognitive stack including topics and hybrid search to a basic 'Agentic' tier for simple navigation.

How does intent classification improve search?

By identifying if a query is meant to 'Explore' or 'Locate,' the skill can adjust search parameters like beam width and node depth to return more relevant results faster.

Can I see why a specific result was returned?

Yes, every retrieval operation returns an explainability payload detailing the layers tried, the intent detected, and the execution time for full transparency.

AI Agent Retrieval Policy

Name: AI Agent Retrieval Policy
Author: SpillwaveSolutions

bySpillwaveSolutions

•

Gestión de Bases de Datos

Optimizes agent memory queries through intelligent intent classification, tiered search routing, and automated fallback chains.

The Retrieval Policy skill serves as the decision-making engine for agent memory systems, determining the most effective strategy for retrieving stored information. It analyzes query intent—distinguishing between broad exploration and specific fact-finding—to route requests through an optimal hierarchy of BM25 keyword matching, vector semantic search, and agentic navigation. By providing automatic fallback mechanisms and detailed explainability payloads, it ensures that AI agents can reliably access context while giving developers full transparency into why specific search methods were chosen.

Características Principales

01Intelligent Query Intent Classification (Explore, Answer, Locate, Time-boxed)

02Configurable Stop Conditions including depth, timeout, and confidence thresholds

03Detailed Explainability Payloads for search transparency and debugging

04Automated Fallback Chains for high-reliability memory retrieval

054 GitHub stars

06Tiered Capability Detection across Topic, Hybrid, and Semantic layers

Casos de Uso

01Routing natural language queries to the most efficient search layer automatically.

02Implementing robust memory retrieval for autonomous agents with fail-safe mechanisms.

03Debugging complex search logic with execution reports and confidence scoring.

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