Stephen

Stephen is the founder of The Prompt Index. He has a background in sales, data analysis and artificial intelligence, and built the site as a free place to find prompts, agent skills and AI tools. The site's guides on AI agent skills, loop engineering, LLM tokens and prompting are published under his name.

Founder, The Prompt Index

• LinkedIn
• X (@Stephen_GPT)
• Contact

Guides by Stephen

A selection of the most-read guides. The blog and blog archive have everything else.

• AI Loop Engineering & Gauntlet Loops (2026)

  Learn AI loop engineering with objective, metric, and boundary design. See how Matt Shumer's Gauntlet Loop works, with templates for code, books, and websites.

  29 Jul 2026

• AI Agent Skills Guide 2026: Directories, Install & Security

  What AI agent skills are, where to find them (directories, registries, platforms), how to install them in Claude Code, Codex and OpenClaw, and keep them safe.

  18 Mar 2026

• SKILL.md Format & Manifest Spec: YAML Frontmatter, Markdown Body, Versioning

  SKILL.md manifest spec for Anthropic and OpenAI Codex skills: YAML frontmatter fields, Markdown body, versioning with metadata.version, and a copyable template.

  19 Apr 2026

• Claude Code vs OpenAI Codex Skills: SKILL.md Frontmatter Compared

  Claude Code vs OpenAI Codex skills: SKILL.md frontmatter (paths, disable-model-invocation), agents/openai.yaml allow_implicit_invocation, folders, portability.

  19 Apr 2026

• GPT-5.6 (Sol) & Claude Fable 5 Prompting Guide (2026)

  How to prompt GPT-5.6 (Sol) and Claude Fable 5 in 2026: reasoning effort, pro mode, programmatic tool calling, subagents, memory, brevity, safeguards, and copy-paste prompts.

  9 Jul 2026

• What Are LLM Tokens? LLM Tokens Explained (2026 Guide)

  An LLM token is a chunk of text (about 4 characters or 0.75 words) that models read and bill by. See how tokenization, context windows and 2026 pricing work.

  9 May 2026

• Jailbreaking LLMs in 2026: The State of Play

  A 2026 field guide to jailbreaking LLMs and prompt injection: which classic attacks are now patched, what still works against AI agents, and how to protect your chatbot.

  14 May 2026

About the site

The Prompt Index is a free-to-browse library of prompts, agent skills, image prompts and GPTs, plus 25 AI tools. Read more on the about page.

The Prompt Index uses AI models in most of its tools. If you spot an error in a guide, please tell us so it can be corrected.