Pre-Post Tipster Claim Stress-Test for Marketplace Compliance Leads

A pre-send gate that intercepts each tipster claim before it posts to a marketplace's subscriber feed. The compliance lead enters or syncs the tipster's structured claim (selection, implied odds, stated confidence, reasoning text). AE's adversarial multi-model debate stress-tests it against the tipster's prior-claim trace, event-context anomalies, and known fraud rhetoric patterns (parlay-stuffing, sample-cherry-picking, post-hoc rationalisation). Output is a structured challenge: which prior claim of this tipster failed under similar context, what disconfirming evidence is absent, how implied confidence diverges from realised hit-rate. After each event resolves against the public results feed, every claim is graded automatically; per-tipster reliability traces accumulate as the compliance lead's regulatory-defence record. AE is uniquely suited because the 6-pattern autopsy taxonomy (Cosmetic Confidence, Concession Laundering, Epistemological Shielding) maps almost line-for-line onto tipster rhetoric failure modes; adversarial multi-model debate produces challenges no single model surfaces; and the <24h grading loop matches event-resolution cadence.

AE's autopsy taxonomy, adversarial debate, and fast reality-graded loop are unusually well-fitted to a narrow, budget-holding compliance buyer at marketplace venues now under explicit post-launch integrity pressure.

Killed: move_cap_reached.