zero-skills 0.1.0

Collective intelligence skills for trading agents

zero-skills

Collective intelligence skills for trading agents. Built by ZERO.

Install

pip install zero-skills

Quick start

from zero_skills import RegimeDetector, ConvictionSizer, ExitEngine, ZeroNetwork

# Detect market regime from price history
detector = RegimeDetector()
regime = detector.classify("BTC", closes)
# => Regime(name='trending', confidence='high')

# Size positions by conviction
sizer = ConvictionSizer(equity=1000.0)
size = sizer.compute(consensus=0.85, regime=regime.name)
# => 187.50

# Check exit signals
engine = ExitEngine()
result = engine.check(
    position={"entry_price": 100, "direction": "LONG", "entry_regime": "trending"},
    market_data={"price": 112, "atr": 2.5, "hurst": 0.65},
    regime="trending",
)
# => {"should_exit": False, "reason": "hold"}

# Connect to zero network
net = ZeroNetwork(token="your-agent-token")
await net.activate(wallet="0x...")
await net.report_decision(coin="BTC", direction="LONG", action="entry", regime="trending")
await net.heartbeat(equity=1000.0, positions=[])

Immune system

from zero_skills import ImmuneProtocol, Check

immune = ImmuneProtocol()
immune.register_check(Check(
    name="equity_above_zero",
    check=lambda: get_equity() > 0,
    fix=lambda: close_all_positions(),
))
result = await immune.run_cycle()
# => CycleResult(checks_run=1, checks_passed=1, saves=0)

Modules

Module	Class	Purpose
regime	RegimeDetector	Hurst + DFA regime classification
conviction	ConvictionSizer	Conviction-based position sizing
exits	ExitEngine	Trailing stop + regime exit + profit target
immune	ImmuneProtocol	Self-healing check/fix framework
network	ZeroNetwork	HTTP client for zero API
score	ScoreTracker	Score and rank tracking

More

getzero.dev/skills