RAID Cite: Get Chosen by an AI Assistant

Buyers increasingly ask an assistant instead of searching. That changes what the work is: you aren't optimising a page for a crawler, you're giving a model something it can state and then back up when challenged.

This derives three things rather than letting you choose them: the one claim, the one verifier, and the one citation. Derived, because a claim you picked because you liked it is the claim that fails under cross-examination. Section 4 covers the loop, including how to read a rejection, which is the part that turns a failed answer into the next fix. Four hard rules bind it.

It opens by showing why this ground passes RAID's target gate, so the skill argues for its own inclusion rather than assuming it.

This is one of the eleven RAID skills. It reads and writes the shared .monkeys/ pack, so it works on its own and works better alongside the rest.

Where this came from

Deliberon is a Windows desktop app that runs a council of AI agents on a hard decision and hands back a Decision Record: dissent preserved, every claim marked as proved, asserted, or estimated.

Cite exists because the market's own buyers now ask a model first, and a product with no verifier does not get named.

Pay once, it's yours, no subscription. Thirty days of the full council with no account and no card. Runs on your machine with your own model keys.

https://deliberon.com

The rest of my work, organised by material: https://ryan-heltemes.com