prompt-architect

Design, audit, and migrate prompts for Claude Opus/Sonnet 4.5/4.6 models.

Updated Mar 21, 2026
One-click install
npx skills add https://github.com/ildunari/open-office-hybrid --skill prompt-architect-ildunari
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: prompt-architect
Source: https://github.com/ildunari/open-office-hybrid/tree/main/skills/prompt-architect
Command: npx skills add https://github.com/ildunari/open-office-hybrid --skill prompt-architect-ildunari

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides a structured methodology for designing, auditing, migrating, and optimizing prompts tailored to Anthropic's Claude models (Opus/Sonnet 4.5/4.6). It helps users craft system prompts, CLAUDE.md files, project instructions, and task prompts while offering robust prompts-quality workflows for prompt audits and migrations.

Core Features & Use Cases

  • Prompt creation and calibration tailored to Claude 4.5/4.6 families, including explicit conditionals, framing choices, and XML format indicators.
  • Audit and migrate existing prompts to align with current model behavior, references, and thinking configurations.
  • Project instruction patterns to persist across chats, reducing drift in long conversations.
  • Use cases include building a new Claude-driven workflow, upgrading an older CLAUDE.md, or performing a migration from Claude 3.5/Opus 4.1 era prompts.

Quick Start

Load calibration references and draft a Claude-ready system prompt and CLAUDE.md using explained conditional guidelines.

Frequently Asked Questions about prompt-architect

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

FAQPage Schema
How do I calibrate system prompts for Claude Opus and Sonnet 4.5/4.6 models?

To calibrate system prompts for Claude Opus and Sonnet 4.5/4.6, apply a structured methodology using explained conditionals, positive framing, and XML format indicators to minimize drift and improve model compliance.

What is the best way to migrate Claude 3.5 prompts to current model behavior?

Migrating Claude 3.5 prompts to current model behavior involves auditing existing instructions and updating them with explicit conditionals, positive framing, and XML format indicators aligning with Opus/Sonnet 4.5/4.6 references.

How do I write a CLAUDE.md file that persists project instructions across chats?

Writing a CLAUDE.md file that persists across chats requires using project instruction patterns with explained conditionals and positive framing to reduce behavioral drift in long conversations.

Why does my Claude prompt drift or ignore instructions in long conversations?

Claude prompts drift in long conversations when lacking structured design. Using project instruction patterns with explicit conditionals, positive framing, and XML format indicators minimizes drift and maintains compliance.

Can I audit existing Claude prompts for Opus and Sonnet 4.5/4.6 compatibility?

You can audit existing Claude prompts for Opus and Sonnet 4.5/4.6 compatibility by running them through a prompt-quality workflow that checks alignment with current references, thinking configurations, and framing guidelines.

When do I need to use XML format indicators in Claude system prompts?

You need XML format indicators in Claude system prompts when defining structured instructions; they help minimize behavioral drift and improve compliance with current Opus/Sonnet 4.5/4.6 model expectations.