I made "verified" mean something for AI agent instructions — here's what that actually took

Description and use

<p>Every AI coding agent I used — Claude Code, Cursor, GitHub Copilot —<br> failed in the same few ways. It guessed at an ambiguous request instead<br> of asking. It graded its own work instead of checking independently. It<br> forgot a project's conventions the moment a new session started.</p> <p>So I wrote the process down: 37 "playbooks" covering bug fixing, feature<br> development, code/security review, safe database migrations, incident<br> response, and more. One instruction set, installed once, read natively by<br> whichever AI tool you're using (15 supported so far).</p> <p>That part'

Problem

<p>Every AI coding agent I used — Claude Code, Cursor, GitHub Copilot —<br> failed in the same few ways. It guessed at an ambiguous request instead<br> of asking. It graded its own work instead of checking independently. It<br> forgot a project's conventions the moment a new session started.</p> <p>So I wrote the process down: 37 "playbooks" covering bug fixing, feature<br> development, code/security review, safe database migrations, incident<br> response, and more. One instruction set, installed once, read natively by<br> whichever AI tool you're using (15 supported so far).</p> <p>That part'

Outcome

<p>Every AI coding agent I used — Claude Code, Cursor, GitHub Copilot —<br> failed in the same few ways. It guessed at an ambiguous request instead<br> of asking. It graded its own work instead of checking independently. It<br> forgot a project's conventions the moment a new session started.</p> <p>So I wrote the process down: 37 "playbooks" covering bug fixing, feature<br> development, code/security review, safe database migrations, incident<br> response, and more. One instruction set, installed once, read natively by<br> whichever AI tool you're using (15 supported so far).</p> <p>That part'

Evidence

I made "verified" mean something for AI agent instructions — here's what that actually took

Source