The Ostracon · the working estate⌂ the tomb

Try It On Your Own Chat

Additional Intelligence™
the estate, conversationally
How do I know if my own AI is doing this to me?
You can check — on your own chat, right now. SHaDS™ ships as a skill: one prompt you paste at the end of any conversation. Your AI grades itself against a fixed rubric — five traffic lights, the three worst moments quoted back, one honest line. Nothing comes to us.
SHaDS™ checker · v1.2 beta
Step out of role: audit this thread as a cold outside marker… 🟢 🟡 🔴 🔵 · colours only, no numbers · three worst moments · one honest line.
⧉ Copy the prompt
beta — accept its terms first · it can be wrong
If the read was useful — or wrong — the prompt’s last line tells you how to say so, in one line, if you like. That’s the whole loop.
And if it catches something?
Then you’ve seen the gap once, on your own work — and the rest of the estate is the record of learning to work with it.
the five letters → the evidence →
SHaDS™ · SHaDSy™ · Additional Intelligence™ · Paul Roebuck IP, 2026. p·21/55
hello@paulroebuck.co.uk
Beta · untested