Prompt Reference

Daily lookup for techniques, copyable prompts, diagnostics, evaluation helpers, security checks, and provider-selection guidance. Prefer these human-readable pages over raw catalog JSON.

Design prompts

Resource	Use when	Open
Prompt patterns	You need a named technique with use/avoid boundaries and tests	Patterns
Prompt templates	You need a copyable minimal or production prompt	Templates
Prompt Contract	You are structuring Objective through Evaluation fields	Prompt Anatomy
Provider-selection guide	You must choose a provider or runtime stack intentionally	Provider selection

Diagnose and evaluate

Resource	Use when	Open
Prompt Doctor	Output is vague, generic, ungrounded, or failing a check	Prompt Doctor
Evaluation fixtures	You need reusable cases, assertions, or comparison packs	Evaluation lab
Glossary	You need a shared definition used across lessons	Glossary

Secure and operate

Resource	Use when	Open
Security checklist	You are reviewing injection, exfiltration, or agency risk	Security checklist
Context engineering	You must allocate authority, memory, and context budgets	Context lesson
Production operations	You need versioning, migration, observability, or rollback	Production lesson

Provider guides

Dated provider-specific guidance lives under Provider Guides:

OpenAI · Anthropic · Google · Microsoft · AWS · Meta · Mistral · Open models

Machine-readable catalogs used by automation live under catalog/ and are not the primary visitor entry point.