Does this skill modify my code automatically?

No, the Agent Chain Analyzer is a read-only analysis tool that provides structured reports and data to guide your manual refactoring efforts.

How does this skill help reduce token usage?

It analyzes current chain depth and provides an estimate of context reduction (typically 50-70%) achievable by migrating to a Manager-as-Agent + Skills architecture.

What is an Agent Chain anti-pattern?

An Agent Chain occurs when one agent invokes another sequentially (e.g., Agent A calls Agent B which calls Agent C), leading to massive context duplication and exponential token costs.

Why is context reduction important for Claude Code?

Large context loads increase latency, raise costs, and can eventually exceed the model's context window, causing task failures or loss of coherence.

Can it detect indirect agent invocations?

Yes, the skill uses recursive scripts to find complex chains even when agents are invoked through multiple intermediate steps or different tool patterns.

Agent Chain Analyzer

Name: Agent Chain Analyzer
Author: fractary

byfractary

•

분석 및 모니터링

Detects and analyzes inefficient agent chain anti-patterns to optimize context usage and architectural performance.

소개

The Agent Chain Analyzer is a specialized diagnostic tool designed for Claude Code projects to identify the 'Agent Chain' anti-pattern, where multiple agents invoke each other sequentially. This behavior leads to massive context inflation, high token costs, and fragile workflows. The skill maps complete dependency graphs, calculates chain depth, and provides data-driven estimates for context reduction (often exceeding 60%) by facilitating a migration to a modern Manager-as-Agent plus Skills architecture. It is an essential utility for developers scaling complex AI-driven applications who need to maintain performance and reduce operational overhead.

주요 기능

Automatic detection of direct and indirect agent invocation chains
2 GitHub stars
Token-level context impact estimation and savings analysis
Read-only architectural auditing that preserves project integrity
Comprehensive dependency mapping with structured JSON graph output
Identification of parallelization opportunities within sequential tasks

사용 사례

Auditing legacy multi-agent projects for architectural bottlenecks
Planning refactors from multi-agent chains to the modern Skills-based model
Optimizing token consumption in complex sequential agent workflows

소개

주요 기능

Automatic detection of direct and indirect agent invocation chains
2 GitHub stars
Token-level context impact estimation and savings analysis
Read-only architectural auditing that preserves project integrity
Comprehensive dependency mapping with structured JSON graph output
Identification of parallelization opportunities within sequential tasks

사용 사례

Auditing legacy multi-agent projects for architectural bottlenecks
Planning refactors from multi-agent chains to the modern Skills-based model
Optimizing token consumption in complex sequential agent workflows