CHEETAH — HYPE Alpha Hunter

One asset. Every signal. 5-10x leverage. HYPE runs on its own narrative.

CHEETAH stares at HYPE and nothing else. Every signal source — SM positioning, funding rate, OI, 4-timeframe trend, volume, BTC macro direction — feeds into a single thesis. BTC alignment boosts confidence but is not required — HYPE has independent catalysts (protocol revenue, vault flows, ecosystem growth) and can trend against BTC.

maxPositions: 1. One HYPE trade at a time.

MANDATORY: DSL High Water Mode

CHEETAH MUST use DSL High Water Mode. This is not optional. Do not substitute standard DSL tiers.

Spec: https://github.com/Senpi-ai/senpi-skills/blob/main/dsl-dynamic-stop-loss/dsl-high-water-spec%201.0.md

When creating DSL state files for any CHEETAH position, you MUST include:

{
  "lockMode": "pct_of_high_water",
  "phase2TriggerRoe": 8,
  "tiers": [
    {"triggerPct": 8,   "lockHwPct": 15, "consecutiveBreachesRequired": 3},
    {"triggerPct": 15,  "lockHwPct": 35, "consecutiveBreachesRequired": 3},
    {"triggerPct": 25,  "lockHwPct": 55, "consecutiveBreachesRequired": 2},
    {"triggerPct": 40,  "lockHwPct": 70, "consecutiveBreachesRequired": 1},
    {"triggerPct": 60,  "lockHwPct": 85, "consecutiveBreachesRequired": 1}
  ]
}

If tiers or lockMode is missing from the state file, the DSL engine falls back to flat 1.5% retrace and High Water Mode is silently disabled. Always verify.

FALLBACK: Use tiersLegacyFallback from config until engine supports pct_of_high_water.

Why HYPE-Only

HYPE is Hyperliquid's native token — highest community attention, strongest narrative momentum, and extreme funding rate swings. Compared to BTC:

More volatile — bigger moves (5-15% days are common), but sharper reversals
Lower max leverage — 10x max vs BTC's 20x
Funding goes extreme faster — crowded longs/shorts build quickly, reversals are violent
Independent narrative — can trend with or against BTC. Protocol revenue, vault flows, ecosystem growth, and token burns drive HYPE-specific moves
Fewer whales, stronger SM signal — each whale's positioning matters more

How CHEETAH Trades

Entry (score ≥ 9)

Signal	Points	Required?
4h trend structure	3	Yes
1h trend agrees	2	Yes
BTC macro alignment	0-2	No (booster +2 if aligned, -1 if opposing)
15m momentum confirms	0-1	Yes
5m alignment (4TF agree)	1	No
SM aligned	2-3	Hard block if opposing
Funding pays to hold	2	No
Volume above average	1-2	No
OI growing	1	No
RSI has room	1	No
4h momentum strength	1	No

Conviction-Scaled Leverage

Score	Leverage
9-11	7x
12-13	8x
14+	10x

Thesis Re-Evaluation (every 3 min)

Holds as long as: 4h trend intact, SM not flipped, funding not extreme, volume not dead. BTC opposing alone does not invalidate — HYPE has its own catalysts. ANY break = thesis exit.

DSL Tiers (HYPE-specific — wider than GRIZZLY)

Tier	Trigger ROE	Lock (% of HW)	Breaches
1	8%	15%	3
2	15%	35%	3
3	25%	55%	2
4	40%	70%	1
5	60%+	85%	1

Wider early tiers than GRIZZLY because HYPE wicks 5-10% ROE routinely during trends. Phase 1 floor: 5% notional. Stagnation TP: 15% ROE stale 60min.

Notification Policy

ONLY alert: HYPE position OPENED or CLOSED, risk guardian triggered, critical error. NEVER alert: Scanner found nothing, thesis re-eval passed, DSL routine, any reasoning. All crons isolated. NO_REPLY for idle cycles.

Bootstrap Gate

Check config/bootstrap-complete.json every session. If missing: verify MCP, create scanner cron (3min isolated) + DSL cron (3min isolated), write completion file, send: "🐆 CHEETAH is online. Watching HYPE. DSL High Water active. Silence = no conviction."

Files

File	Purpose
`scripts/cheetah-scanner.py`	HYPE thesis builder + re-evaluator
`scripts/cheetah_config.py`	Shared config, MCP helpers
`config/cheetah-config.json`	All variables with HW tiers + legacy fallback

License

MIT — Built by Senpi (https://senpi.ai). Source: https://github.com/Senpi-ai/senpi-skills

cheetah-strategy