exploration-strategy

Define parallel exploration workflows for locating codebase structure and patterns.

7|1|Updated Feb 2, 2026
One-click install
npx skills add https://github.com/josix/agent-flow --skill exploration-strategy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: exploration-strategy
Source: https://github.com/josix/agent-flow/tree/main/skills/exploration-strategy
Command: npx skills add https://github.com/josix/agent-flow --skill exploration-strategy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill solves the problem of wasting time during unfamiliar codebase exploration by providing a structured, convergent approach to finding relevant files, patterns, and conventions before planning implementation.

Core Features & Use Cases

  • Parallel exploration patterns: Run multiple independent searches (glob/grep/read/graph queries) to quickly cover different angles of the repository.
  • Tool selection guidance: Choose between Graphify graph tools for structural questions and grep/glob/read for literal content and freshly edited code.
  • Convergence criteria and handoff summaries: Stop when results stabilize and produce actionable summaries for downstream planning and implementation.

Quick Start

Ask an agent to use exploration-strategy to quickly orient you in a new repository, gather evidence-backed context, and summarize what files to modify next.

Frequently Asked Questions about exploration-strategy

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

FAQPage Schema
What is the best way to explore an unfamiliar codebase before planning changes?

Codebase exploration is fastest when using parallel search workflows that run independent grep, glob, and read queries simultaneously to locate relevant file structures, patterns, and conventions before implementation.

How do I map dependencies in a large repository effectively?

Mapping dependencies requires selecting structural exploration patterns like breadth-first or depth-first searches, utilizing graph tools for structural queries, and chaining search primitives to produce an evidence-based handoff summary.

When should I use graph tools versus grep for codebase context?

Graph tools work best for structural questions and dependency mapping, while grep, glob, and read are more effective for locating literal content, specific patterns, and freshly edited code within the repository.

How to conduct a refactoring impact analysis across a codebase?

Refactoring impact analysis involves running targeted parallel searches across the codebase to map dependencies, identify conventions, and gather convergent evidence that determines which files require modification.

Can I limit parallelism when running multiple codebase searches?

Parallelism should be limited during codebase exploration to maintain signal quality, ensuring the multiple independent search queries produce stable, convergent results rather than overwhelming noise.

What convergence criteria stop codebase exploration workflows?

Codebase exploration stops when search results stabilize and no longer surface new relevant files or patterns, at which point an actionable handoff summary is produced for downstream planning.