How does Agent Architect prevent context drift?

It identifies 'immutable laws' that must be displayed at the start of every response and ensures they are placed in CLAUDE.md with self-reinforcing display triggers.

What are the size limits recommended by Agent Architect?

To maintain efficiency, it recommends keeping Rule files under 50 lines and Skill files under 200 lines, referencing external documentation for deeper details.

What is the main purpose of the Agent Architect skill?

It acts as a strategic advisor that organizes project knowledge into the correct Claude Code components such as laws, skills, commands, and rules.

Does Agent Architect execute code changes directly?

No, it is an advisor. It classifies the task and delegates the execution to specific skills like 'write-skill' or 'write-command' to maintain a clean separation of concerns.

Agent Architect

Name: Agent Architect
Author: wayne930242

bywayne930242

•

生产力与工作流

Architects and organizes Claude Code knowledge by classifying information and delegating tasks to specialized development skills.

The Agent Architect skill acts as a high-level orchestration layer for Claude Code, ensuring that project knowledge, rules, and capabilities are stored in the correct locations to prevent context drift and redundancy. By following a structured classification framework, it distinguishes between immutable laws in CLAUDE.md, operational skills, user commands, and static rules, maintaining a clean and scalable agentic workflow. This skill is essential for developers looking to build complex, multi-agent systems within Claude Code while maintaining a single source of truth and strict modularity.

主要功能

01Structural integrity validation to prevent content duplication

02Context engineering optimization for long-running conversations

03Standardized management of skills, commands, subagents, and hooks

04Automated task delegation to specialized writer skills

054 GitHub stars

06Structured knowledge classification framework (Laws vs. Knowledge)

使用场景

01Auditing existing skills and rules to ensure they adhere to size limits and single-responsibility principles

02Restructuring a Claude Code environment to improve modularity and reduce context noise

03Determining the optimal storage location for new project guidelines or technical constraints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add wayne930242/reflexive-claude-code agent-architect

For use in Claude.ai and ChatGPT

主要功能

01Structural integrity validation to prevent content duplication

02Context engineering optimization for long-running conversations

03Standardized management of skills, commands, subagents, and hooks

04Automated task delegation to specialized writer skills

054 GitHub stars

06Structured knowledge classification framework (Laws vs. Knowledge)

使用场景

01Auditing existing skills and rules to ensure they adhere to size limits and single-responsibility principles

02Restructuring a Claude Code environment to improve modularity and reduce context noise

03Determining the optimal storage location for new project guidelines or technical constraints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add wayne930242/reflexive-claude-code agent-architect

For use in Claude.ai and ChatGPT