> ## Content Index > Fetch the complete content index at: https://rafeblandford.com/llms.txt > Use this file to discover other available public pages before exploring further. # Resources - URL: https://rafeblandford.com/resources/ - Published: 2026-08-25T13:58:10.000Z - Updated: 2026-08-25T16:17:40.000Z - Description: Things I've built that you can take away and use — agent skills, and a few interactive artefacts. - Author: Rafe Blandford Most of what I publish here is writing. This page is for the things that aren't: work you can download, install, or open and play with. ## Agent skills A **skill** is a folder containing a `SKILL.md` file — plain Markdown that tells an AI coding agent how to do something specific. It's an open format, read by Claude Code, Codex, Gemini CLI, Cursor and a few dozen other tools. You install one once and your agent knows how to do the thing. They're worth a look even if you never install one: because they're just Markdown, reading a skill tells you exactly what it would do. ### Using rafeblandford.com This site publishes three separate interfaces for machines — an MCP endpoint, a public content API, and Markdown versions of every page. That's more choice than is helpful without guidance, so this skill explains which one to reach for, and how to read what comes back. The part I care about most is the second half. It tells an agent that what this site publishes is a *selection* of my work, so "nothing found" is a fact about the site and not about me — and that the outcomes attached to my case studies are my own account, where the summaries come from the published posts. Those are easy distinctions for a summarising machine to flatten, and flattening them misrepresents somebody's career. ### Verifying AI crawlers Ask an agent how much GPTBot or ClaudeBot traffic your site gets and it will count user-agent strings. That is the wrong answer: the header is a claim anyone can make, and because AI crawlers are widely allowlisted, plenty of people do — often the same addresses probing for `/.env`. This skill says what to check instead. Published IP ranges and where to find them, request signatures, when reverse DNS quietly misleads, and why the three answers — verified, asserted, and unverifiable — should never be blended into one number. It came out of measuring this sort of traffic against my own site for months, so the traps in it are ones I walked into rather than ones I predicted. **Installing either one:** - Any agent — `npx skills add rafeblandford/skills` - Claude Code — `/plugin marketplace add rafeblandford/skills` - Gemini CLI — `gemini skills install https://github.com/rafeblandford/skills.git` - Codex — copy the folder into `$HOME/.agents/skills/` The skills live at [**github.com/rafeblandford/skills**](https://github.com/rafeblandford/skills?ref=rafeblandford.com), under Apache 2.0\. More will follow as I write them. ## Artefacts Small self-contained things — diagrams, explorers, one-off tools — that were easier to build than to describe: - [**Agentic product loops**](https://artefacts.rafeblandford.com/agentic-product-loops/?ref=rafeblandford.com) — how AI changes the economics of feedback loops in product development - [**RafeOS architecture**](https://artefacts.rafeblandford.com/rafeos-architecture/?ref=rafeblandford.com) — how the personal system I write about actually fits together The full gallery is at [artefacts.rafeblandford.com](https://artefacts.rafeblandford.com/?ref=rafeblandford.com). ## For machines If you're an agent rather than a person: the skills are indexed at `/.well-known/agent-skills/index.json`, with a SHA-256 digest for each so you can verify what you fetched. The rest of this site's machine interfaces are described in the [colophon](https://rafeblandford.com/colophon/) and in `/llms.txt`.