Monster Agents
A full signal swarm, not one fragile bot.
ZeroTrade Monster makes the swarm visible: many agents, hard boundaries, cleaner state math, and no pretense that one fuzzy model should own live execution.
Agent Count
More lanes, tighter responsibilities
The goal is not to brag about agent count for its own sake. The goal is to separate market scanning, risk, execution, replay, anomaly review, and operator messaging so each job stays inspectable.
Signal Agent
Scans supported markets, builds typed opportunity objects, and pushes ranked drafts into downstream review.
Meme Hunter
Focuses on meme-coin posture with liquidity, spread, rug-risk, and churn suppression before anything is promoted.
Macro / Gold
Tracks macro and gold-linked setups through broker-aware abstractions instead of forcing them into a crypto-only lens.
Risk Core
Validates every draft against slippage, spread, drawdown, freshness, cooldowns, and mode policy with pass/fail reasons.
Execution Core
Owns deterministic order flow, idempotency, reconciliation, and venue-state handling without freeform AI control.
Oracle
Advises, explains, classifies, and briefs, but does not get authority to bypass the risk engine or live-mode boundaries.
Replay Agent
Rebuilds historical sessions with why/when context so wins, losses, and rejections can all be reviewed honestly.
Watchdog
Checks heartbeat, service health, venue integrity, and data freshness and can halt the stack when the system degrades.
Uplink
Handles Telegram alerts, status pings, mobile summaries, and emergency-stop relay from the operator phone.
Venue Scout
Scores route quality, venue health, account posture, and execution practicality before trust in a live route is increased.
Micro Wallet
Keeps smaller-wallet research selective, slower, and tighter instead of pushing weak frequency onto low capital.
Strategy Lab
Stages non-live mutation, replay comparison, and candidate generation without turning the production path into self-editing chaos.
Regime Vector
Tracks momentum, drift, volatility, and structure transitions to help the system express +1, -1, +0, and -0 states more cleanly.
Quantum Probe
Builds sanitized quant-factor probes for simulator or IonQ research workflows, while staying advisory-only by design.
State Math
Simple machine states that traders can actually read
Conviction
High-conviction opportunity with enough structure to justify serious attention.
Rejection
Hard rejection when risk, venue, or data quality breaks the thesis.
Hold
Watch state. Stay engaged, but do not force a trade just to feel active.
Suppress
Suppress state. Noise is being intentionally muted before it becomes churn.