iterative-retrieval

Iteratively retrieve codebase information using Glob, Grep, and AST searches.

37|6|Updated Feb 14, 2026
One-click install
npx skills add https://github.com/majiayu000/vibeguard --skill iterative-retrieval-majiayu000
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: iterative-retrieval
Source: https://github.com/majiayu000/vibeguard/tree/main/skills/iterative-retrieval
Command: npx skills add https://github.com/majiayu000/vibeguard --skill iterative-retrieval-majiayu000

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill addresses the challenge of finding specific information within large codebases by employing a multi-stage, iterative search process that refines queries over time.

Core Features & Use Cases

  • Multi-stage Search: Utilizes a 4-phase loop (DISPATCH, EVALUATE, REFINE, LOOP) to progressively narrow down search results.
  • Flexible Search Strategies: Supports Glob (filename pattern), Grep (content keyword), and AST (code structure) searches.
  • Use Case: When tasked with understanding a complex feature, use this Skill to first broadly search for relevant files, then evaluate the initial results to refine keywords and search locations, and finally loop until precise code snippets or relevant documentation are identified.

Quick Start

Use the iterative retrieval skill to find all Python files related to user authentication.

Frequently Asked Questions about iterative-retrieval

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

FAQPage Schema
How do I locate specific code snippets in a large codebase when initial search results are too broad?

Iterative codebase search narrows down broad results by using a four-stage loop: DISPATCH, EVALUATE, REFINE, and LOOP. It progressively refines search queries over multiple iterations to pinpoint exact code snippets or documentation.

What search strategies can I use for codebase navigation to find files and structural code?

Codebase navigation supports three search strategies: Glob for filename patterns, Grep for content keywords, and AST for code structure. These strategies can be combined across iterations to locate relevant files and precise code implementations.

How does iterative information retrieval work when searching for complex features?

Iterative information retrieval works by first broadly searching for relevant files, evaluating initial results to refine keywords and locations, then looping up to three rounds until precise code or documentation is identified.

Can I use this approach to identify existing implementations and flag duplicate code?

Yes, the iterative search process integrates with VibeGuard to identify existing implementations and flag duplicate code. This helps prevent redundancy while precisely locating relevant code across the codebase.

What are the limitations of using a multi-stage search loop for code search?

The multi-stage search loop is limited to a maximum of three refinement rounds. This constraint bounds the iterative retrieval process, requiring effective query refinement within the EVALUATE and REFINE phases to achieve precise results.

Does iterative code search require any specific dependencies or environment setup?

No specific dependencies or environment components are required to run iterative code search. The skill operates independently using its built-in search strategies and VibeGuard integration to navigate codebases.