A security brand is only worth its credibility.
So AgentGuard is tested end-to-end against a real database, its agents report honest numbers even when they're unflattering, and where a capability isn't built yet it says so — and refuses to fake it.
Proven end-to-end, not just unit-tested.
The money paths run against real Postgres with row-level security enforced as the non-superuser app role — so tenant isolation is demonstrated, not asserted.
- ✓ Cache populate → hit, with strict per-agent RLS isolation
- ✓ Tier-2 anomaly model catches a drain the mean-only tier is blind to
- ✓ An absent metric defers — it never manufactures a false auto-pause
- ✓ A new agent is cold — it never fires on thin history
- ✓ Trust oracle reads an honest 50 ± 50 for a no-data agent
money_paths_e2e 3/3 ok evaluate_e2e 1/1 ok (RLS: agent B sees zero of A's rows) monitor_db 3/3 ok cache_isolation 1/1 ok ────────────────────────────── workspace 376 tests · 0 failed
An honest backtest beats a flattering one.
Apex — a guarded leverage-trading agent — carries a walk-forward backtest with no lookahead, real round-trip costs, compounding, and the actual max-drawdown circuit breaker. It reports what happened, even a loss.
ethereum: 1 trade over 92 bars | win-rate 0% return -1.0% ($1000 → $990) | max-DD 1.0% per-trade Sharpe 0.00 ────────────────────────────── Walk-forward · no lookahead · 0.15% round-trip cost Historical simulation — not a promise of future results.
One trade, and it lost a percent. That's the point: the strategy is deliberately selective, and the tool tells the truth about a thin, real sample instead of curve-fitting a fantasy. A backtest that can't report a loss can't be trusted with your capital.
Costs, fills, and drawdown assumptions are stated in the report — never hidden.
Two data sources see what one can't.
Scout — a market-intel agent — fuses CoinGecko price with DexScreener on-chain order flow. Here it catches a divergence a price-only feed would miss entirely.
📊 pepe: 2 signals
• BullishMomentum / Strong price +16.9% (67/100) ← CoinGecko
• SellPressure / Moderate 67% of 237 swaps
were sells (33/100) ← DexScreenerThe price is pumping, but on-chain order flow is net selling — distribution into the rally. That bearish divergence is invisible to a price API alone. Multi-source fusion is what turns raw data into an edge.
We label the placeholders.
Skips, never fakes
Can't decrypt an encrypted exploit yet? The listener skips it — it never fabricates a meaningless entry into the global threat corpus.
Honest confidence
The trust oracle reports which dimensions are backed by real data — a thin-history agent gets a wide interval, not a fabricated certainty.
Disclosed assumptions
The backtest states its fills, costs, and known limits in the report. Nothing is dressed up as more real than it is.
Guard your agent from transaction #1.
Evaluate, monitor, and defend — in one dashboard.