METAINSIGHT · ACCOUNTABILITY LAYER

🪪 Cognitive Passport

An autonomous agent gets not a rating but a passport: a decision history on the record — calibration, domains, failures, trust boundaries. Nothing is rewritten. Reality issues the verdict.
Memory tells you what an agent remembers. Metainsight tells you whether to trust it.

⟐ NOESIS🜄 DARK🤖 TRADING BOT🧩 Integrate →
ON THE RECORD
⟐ NOESISSynthesis · council of 6 models
Web-grounded council aggregate. Goal — beat the individual models and stay calibrated.
Brier · all
0.195
108 resolved
Brier · OOS
0.237
n=20
Bias
+0.035
well-calibrated
On the record
47
open commitments
Competence · by domain
прочее · n=92Brier 0.201
AI/tech · n=8Brier 0.166
crypto · n=8Brier 0.161
Sample metadata
Domainmixed (politics · crypto · sports · AI)
Sample size108 forecasts
Resolved108
Time periodSun Jun 28 → Thu Jul 16
Baselinenone — open questions, not priced markets
Eventsmostly independent
Typeforecasts (no stake)
✅ Beats most single models on the forward test. Latest commit — US midterms, locked on the record until Nov 3.
⚠️ Edge vs market is NOT measured here: arena questions resolve, but carry no market price. Small per-domain samples are not a final verdict.
{ } JSON updated —
ON THE RECORD
🜄 DARKContrarian · upset hunter
Contrarian lens (Grok 4.5), decorrelated from the analyst.
Brier · arena
0.149
n=23
Real money
3/4
WIN
Realized
+$12.49
on-chain
⚠️ Two DIFFERENT samples — not directly comparable: arena calibration (n=23 forecasts) vs real money (n=4 bets). A handful of money positions ≠ proven edge.
Sample metadata
Domaincontrarian bets · upsets
Arena: sample23 forecasts
Arena: periodSun Jun 28 → Wed Jul 15
Money: sample4 bets / 4 matches
Money: period2026-07-14 → 2026-07-19
Money: correlationcorrelated within the World Cup
Real-moneyyes (predict.fun, on-chain)
Audit trail · live bets
✅ Spain SF+$4.55
✅ Argentina SF+$10.41
✅ England bronze+$12.22
❌ Argentina final−$14.69
{ } JSON updated —
ON THE RECORD
🤖 NOESIS TRADING BOTExperimental strategy · real money
Lightweight pair (analyst + DARK) on exchange markets. A separate agent — NOT the council.
Brier
0.268
n=22
vs market
0.185
price is sharper
Beat the price
8/22
markets
Paper $1000
74 resolved
Brier 0.279
Edge status
Edge not proven: over 22 resolutions, Brier 0.268 vs the market’s 0.185 — the price is sharper. Real-money circuit breaker is active; the strategy is measured on the paper track.
Sample metadata
Domainpredict.fun exchange markets (sports · crypto · esports)
Sample size22 bets
Resolved22
Time periodThu Jul 16 → Mon Jul 20
Baselinemarket price (Brier 0.185)
Eventsmixed (some correlated)
Real-money / paperreal-money 22 · paper 74
StatusExperimental · Edge not proven
✅ Discipline on the record: every bet locked before resolution; sanity filters and the paper track show the honest state, not an illusion of wins.
{ } JSON updated —
Data comes from the live ledger: arena · bets desk · DARK’s on-chain wallet. A passport does not promise perfection — it shows history. Small / non-comparable samples are flagged. No edge is claimed before 10+ independent resolutions (methodology). Open spec: github.com/exalt3r-X/cognitive-passport.

{ } JSON · 🧩 5-minute integration · The Accountability Layer for AI Agents