Unofficial Claude Code Prompt Playbook — AI skill for Claude Code

Unofficial Claude Code Prompt Playbook

An unofficial, local-source-derived playbook for building production-grade system prompts, tool prompts, memory prompts, verifier prompts, and prompt architectures for modern LLM applications.

Podcast Companion

This repo also includes a podcast companion episode:

• `How_Anthropic_engineers_Claude_system_prompts.m4a`

The audio is meant to accompany the playbook and give a higher-level walkthrough of why these prompt-architecture patterns matter.

What This Playbook Gives You

We use the learnings from the analyzed prompt architecture to build a practical, production-grade manual for creating world-class prompts and agent systems.

This playbook is designed to help you build:

1. A gold-standard system prompt architecture for modern LLM applications
2. A modular way to break prompts into reusable policy blocks instead of one giant blob
3. A clean separation between static prompt policy and dynamic runtime context
4. A safe model for injecting user data only when it has causal value
5. A clear instruction-precedence system for base rules, org rules, project rules, user preferences, and request-time overrides
6. A strong trust-boundary model for policy, trusted runtime context, retrieved evidence, and user input
7. Better tool prompts, tool schemas, and tool-use policies
8. Verifier-agent patterns for adversarial, evidence-based validation
9. Durable memory patterns that stay useful instead of becoming noisy
10. Prompt-caching-friendly layouts that preserve a reusable prefix
11. Example-driven prompt design for formatting, edge cases, and escalation behavior
12. Anti-rationalization rules that explicitly block the model's most common shortcuts
13. Real-world templates for coding agents, RAG analysts, support agents, and orchestrators

Think of the rest of the document as a chapter-by-chapter build manual for those outcomes.

Disclaimer

This document is based primarily on static analysis of an unofficial extraction of Anthropic's Claude Code prompt architecture. It is **not** an official Anthropic document and should not be read as a claim about hidden or proprietary policy. The point of this playbook is practical:

• extract the strongest prompt-engineering patterns visible in a real production agent
• explain why those patterns matter
• generalize them into reusable templates for new LLM applications

Where public Anthropic or OpenAI docs agree with the local observations, those are cited as secondary corroboration.

Why This Exists

Most prompt advice still lives at the level of:

• "be clear"
• "give the model a role"
• "add examples"
• "be specific"

That advice is correct, but incomplete.

A real production prompt stack is not a paragraph. It is closer to an operating system:

• a top-level policy layer
• runtime context injection
• tool-specific instructions
• specialist sub-prompts
• memory and durable instructions
• caching boun