How many cycles does the retrieval loop typically perform?

The pattern is designed to run for a maximum of 3 cycles, which is usually sufficient to identify high-relevance context while maintaining high performance.

What is the Iterative Retrieval pattern?

It is a systematic 4-step loop—Dispatch, Evaluate, Refine, and Loop—designed to help AI agents discover the most relevant codebase context without exceeding token limits.

How does it prevent 'context overflow'?

Instead of sending the entire codebase to an agent, it starts with broad queries and narrows down to only the most relevant files based on a relevance scoring system.

Does this skill work with any programming language?

Yes, the pattern is language-agnostic. it relies on keyword matching, file pattern discovery, and semantic relevance evaluation regardless of the source language.

Iterative Context Retrieval

Name: Iterative Context Retrieval
Author: affaan-m

byaffaan-m

•

156,032

•

Productivity & Workflow

Optimizes AI agent context by incrementally searching and refining code retrieval through a multi-step evaluation loop.

Iterative Context Retrieval solves the common 'context problem' in multi-agent workflows where sub-agents often lack the necessary codebase knowledge to perform specialized tasks accurately. By employing a structured four-stage process—Dispatch, Evaluate, Refine, and Loop—this skill allows agents to start with broad intent and systematically narrow down the most relevant files. This prevents context window overflows and reduces 'missing context' failures, ensuring Claude has the precise implementation details needed for complex bug fixes or new feature development without wasting tokens on irrelevant data.

Key Features

01Explicit gap identification to drive search refinement

02Automated token usage optimization for sub-agents

03156,032 GitHub stars

04Relevance scoring system (0.0 to 1.0) for context quality

05Dynamic codebase terminology and pattern learning

06Four-stage retrieval loop (Dispatch, Evaluate, Refine, Loop)

Use Cases

01Gathering precise context for complex bug fixes in unfamiliar repositories

02Discovering specific project patterns and middleware for new feature implementation

03Optimizing agent orchestration to prevent context window limit failures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code iterative-retrieval

For use in Claude.ai and ChatGPT

Key Features

01Explicit gap identification to drive search refinement

02Automated token usage optimization for sub-agents

03156,032 GitHub stars

04Relevance scoring system (0.0 to 1.0) for context quality

05Dynamic codebase terminology and pattern learning

06Four-stage retrieval loop (Dispatch, Evaluate, Refine, Loop)

Use Cases

01Gathering precise context for complex bug fixes in unfamiliar repositories

02Discovering specific project patterns and middleware for new feature implementation

03Optimizing agent orchestration to prevent context window limit failures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code iterative-retrieval

For use in Claude.ai and ChatGPT