Layer status
- Trust Ledger active
- Action Ladder active
- Paper-to-Live active
- Strategy Passport active
Experiment decision · observed 2026-07-30
AI Trader signal-copying — stopped
The external crowd did not show a tradeable edge in our independent forward evaluation. BibaMoney preserves the failed result and moves the product toward reproducible strategy audits.
GOGA safety layers · public read-only
Safety & governance snapshot
Aggregated from Tvijo AIOS /api/v1/trading/safety/*. No user accounts, ledger, confirmations, or execution controls on this surface.
Action Ladder
- Read
- 1
- Explain
- 2
- Simulate
- 2
- CONFIGURE
- 1
- LIVE_EXECUTE
- 1
- PAPER_EXECUTE
- 1
- SPEND
- 1
Strategy Passport
- Passports
- 10
- Decaying
- 0
- backtested · 10
The decision in four numbers
Real Gate.io 1h candles, next-bar entry, 5 bps fee each side, no leverage.
15,501
749 agents · 427 symbols
−0.3733%
After round-trip fees
−0.6988%
Worse than the same market window
45.1%
Not enough to offset losses and fees
What is stopped
The old prompt and preset leaderboard were demonstrations, not proof.
- No copying external AI-agent signals into a trading plan.
- No hourly ingest, optimizer cross-reference or outbound signal publication.
- No claim that an LLM directive, backtest rank or synthetic-candle demo can execute safely.
- No capital scaling from this experiment.
What continues
The evidence is useful even when the strategy idea fails.
- Keep the frozen dataset and negative verdict auditable.
- Test submitted strategies with provenance, realistic costs, OOS and null controls.
- Publish repeatable Strategy Passports and clear failure modes.
- Validate paid strategy audits before adding more automation.
The product is the verdict, not the signal
Exact specification, realistic simulation, untouched OOS, uncertainty and a repeatable report can save users from a bad strategy without promising returns.
Read the methodology → Open the signed-in evidence page