Can I resume a deep analysis if it gets interrupted?

Yes, the skill includes built-in checkpointing that saves progress at every phase, allowing you to resume exactly where the analysis left off.

What is the Deep Analysis skill for Claude Code?

It is an advanced workflow that uses a lead agent to coordinate a team of specialized agents to explore, map, and synthesize information about a codebase.

How does the hub-and-spoke coordination work?

A lead agent performs initial reconnaissance and creates a plan, assigning specific focus areas to 'worker' agents who explore independently before a synthesizer merges all findings.

Does this skill use a lot of tokens?

While comprehensive, it includes a configurable caching system (TTL) that prevents redundant exploration of files that haven't changed, significantly optimizing resource usage.

Is user approval required for the analysis plan?

Yes, by default, the skill presents a detailed team plan for user approval before starting any heavy exploration, though this can be toggled in settings.

Deep Analysis Workflow

Name: Deep Analysis Workflow
Author: sequenzia

bysequenzia

•

生産性とワークフロー

Conducts comprehensive codebase explorations and architectural synthesis using coordinated multi-agent teams and dynamic planning.

Deep Analysis is a sophisticated workflow skill for Claude Code designed to automate the heavy lifting of codebase investigation. By orchestrating a 'hub-and-spoke' team of agents, it performs rapid reconnaissance to map project structures, identifies high-impact focus areas, and coordinates parallel exploration tasks. This skill is ideal for developers onboarding to legacy systems, performing impact analysis for major refactors, or generating deep architectural insights. It features robust session management including smart caching, progress indicators, and checkpointing to ensure reliable results in complex engineering environments.

主な機能

01Dynamic reconnaissance-based planning tailored to specific project structures

02Smart caching of exploration results to optimize token usage and speed

03Multi-agent team orchestration for parallel codebase exploration

04Session checkpointing and recovery for long-running analysis tasks

05Configurable approval gates for team plans and resource allocation

0622 GitHub stars

ユースケース

01Onboarding to large, unfamiliar codebases to quickly understand architecture and patterns

02Performing comprehensive impact analysis before implementing major features or refactors

03Synthesizing deep technical documentation for complex, undocumented repositories

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sequenzia/agent-alchemy deep-analysis

For use in Claude.ai and ChatGPT

Download Skill