iterative-retrieval

Refine file retrieval across DISPATCH, EVALUATE, REFINE, and LOOP phases.

Updated Jan 15, 2026
One-click install
npx skills add https://github.com/sori883/certification-proflie --skill iterative-retrieval-sori883
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: iterative-retrieval
Source: https://github.com/sori883/certification-proflie/tree/main/.claude/skills/iterative-retrieval
Command: npx skills add https://github.com/sori883/certification-proflie --skill iterative-retrieval-sori883

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Addresses context gaps in multi-agent workflows by progressively refining retrieved information to guide sub-agents.

Core Features & Use Cases

Key capabilities include staged DISPATCH, EVALUATE, REFINE, and LOOP cycles, relevance scoring, dynamic query refinement, and selective exclusion to build actionable context for complex tasks.

Quick Start

Begin with broad file queries to collect candidates, then iteratively refine search criteria based on relevance until you have sufficient high-quality context.

Frequently Asked Questions about iterative-retrieval

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I resolve context gaps in multi-agent workflows?

Context gaps in multi-agent workflows are resolved through staged, iterative retrieval that progressively discovers codebase patterns and terminology across DISPATCH, EVALUATE, REFINE, and LOOP phases. This automated loop builds actionable context for complex tasks.

What is the best way to refine code search queries for sub-agents?

Refining code search queries for sub-agents is best achieved by starting with a broad file query, scoring relevance, and dynamically refining search criteria with selective exclusion. Up to three cycles surface high-relevance files to guide sub-agents effectively.

When do I need iterative retrieval for codebase context?

Iterative retrieval for codebase context is needed when patterns and terminology in the codebase are unclear for sub-agents. It applies progressive discovery to resolve ambiguity, enabling reliable outcomes through staged context refinement cycles.

How does relevance scoring work in iterative context refinement?

Relevance scoring in iterative context refinement evaluates retrieved file candidates during the EVALUATE phase. It assesses search results against dynamic criteria, allowing the LOOP phase to refine queries and selectively exclude low-relevance files over up to three cycles.

Can I use automated loops for multi-agent context management with large codebases?

Automated loops for multi-agent context management handle large codebases by applying up to three iterative retrieval cycles. The staged process dispatches broad queries, evaluates relevance, and refines criteria to surface high-relevance files effectively.

What are the limitations of iterative retrieval in multi-agent workflows?

Limitations of iterative retrieval in multi-agent workflows include a strict maximum of three cycles to surface high-relevance files. If codebase context remains unclear after these refinement loops, sub-agents may still lack sufficient actionable context for complex tasks.