Polished text is no longer scarce. Judgment under real constraints is.
THE MANDORKA MANIFESTO
Claims are cheap.
Proof is not.
AI can make weak work look polished and strong work move faster. The hiring system needs a better signal: not whether someone can produce output, but whether they can control the quality of AI-assisted work.
WHAT WE BELIEVE
The résumé is a claim. The WorkProof is an observation.
AI use is not the problem. Uncontrolled AI output is the problem.
A hiring signal should show what happened, under which version, with which scoring rules and integrity controls.
No company should need to trust a screenshot when a canonical record can be verified live.
No candidate should need permission from one platform to carry evidence of their own work capability.
No benchmark deserves authority until the underlying measurement has been tested against reality.
The winning standard will be embedded in actual hiring and work workflows, not the one with the loudest certificate.
THE CATEGORY
Control of AI-assisted work.
Mandorka is not trying to prove that a person never uses AI. It measures whether the person notices errors, preserves instructions, verifies claims, prioritizes correctly and communicates responsibly while AI is part of the workflow.
Own the proof.
Build a portable record that can be opened and verified without a recruiter account.
Inspect the evidence.
Compare relevant WorkProofs before expensive interview time and use dimensions for targeted follow-up.
Earn the standard.
Version the instrument, harden integrity, validate the signal and publish limitations instead of manufacturing authority.
THE END STATE
Work capability should be verifiable infrastructure.
When a candidate says “I can control AI-assisted operations work,” the natural next question should become: show the WorkProof.