name: wolverine-strategy
description: >-
WOLVERINE v1.0 — HYPE alpha hunter with position lifecycle. Single asset, every signal
(SM, funding, OI, 4TF trend, volume, BTC correlation). Three-mode lifecycle:
HUNTING (scan for entry) / RIDING (DSL trails) / STALKING (watch for reload on dip).
After DSL takes profit, watches for fresh momentum impulse while confirming macro thesis
is intact. If thesis dies, resets. If dip reloads, re-enters. DSL High Water Mode (mandatory).
license: MIT
metadata:
author: jason-goldberg
version: "1.1"
platform: senpi
exchange: aftermath
base_skill: grizzly-v2.0

WOLVERINE v1.1 — HYPE Alpha Hunter with Position Lifecycle

One asset. Every signal. Maximum conviction. Reload-on-dip.

v1.1 Changes — DSL Recalibrated for HYPE Speed

Day 1 data: Wolverine caught a textbook HYPE LONG — score 10+, SM aligned, peaked at +9.3% ROE (+$18.54). Then gave it ALL back because the DSL was too loose. Price dropped from +9.3% to -2.3% ROE faster than the 3-minute scan could react. The old Tier 1 locked only 20% of high water = 1.86% floor, which was meaningless.

v1.1 fix: Phase 2 trigger lowered from 6% to 3% ROE. Tier 1 lock raised from 20% to 40%. Breaches reduced from 3 to 2. HYPE reverses fast — the DSL must lock profits aggressively.

With v1.1 DSL, that same trade locks +3.7% ROE at worst instead of exiting at -2.3%.

Also in v1.1: BTC correlation is bonus-only, never a gate or thesis exit. HYPE moves independently of BTC — up 50% while BTC dropped 30%. When BTC confirms, HYPE runs harder (+2 score bonus). When BTC diverges, it's a non-event.

WOLVERINE stares at HYPE and nothing else. Every signal source available — smart money positioning, funding rate, open interest, 4-timeframe trend structure, volume, BTC correlation — feeds into a single thesis: is there a high-conviction HYPE trade right now?

Based on GRIZZLY v2.0's three-mode lifecycle, adapted for HYPE's volatility profile.

The Three-Mode Lifecycle

MODE 1 — HUNTING (default)

Scan HYPE every 3 minutes. All signals must align (4h trend, 1h momentum, SM, funding, OI, volume). Score 10+ to enter. When a position opens, switch to MODE 2.

MODE 2 — RIDING

Active position. DSL High Water trails it. Thesis re-evaluation every scan. If thesis breaks (4h trend flips, SM flips, funding extreme, volume dies) -> thesis exit and reset to MODE 1. If DSL closes the position -> switch to MODE 3. Note: BTC divergence does NOT invalidate HYPE thesis — HYPE moves independently.

MODE 3 — STALKING

DSL locked profits. The trend may not be over. Watch for a reload opportunity. Every scan checks:

Reload conditions (ALL must pass):
1. At least one completed 1h candle since exit (~30 min minimum)
2. Fresh 5m momentum impulse in the exit direction
3. OI stable or growing
4. Volume at least 50% of original entry
5. Funding not spiked into crowded territory
6. SM still aligned in the exit direction
7. 4h trend structure still intact

If ALL pass -> RELOAD. Re-enter same direction with dynamic leverage clamp (lev = min(targetLev, market.maxLeverage())). Switch to MODE 2.

Kill conditions (ANY triggers reset to MODE 1):
- 4h trend reversed
- SM flipped against exit direction
- OI collapsed 20%+
- Stalking for more than 6 hours with no reload
- Funding spiked above 100% annualized

maxPositions: 1. WOLVERINE holds one HYPE position at a time.

MANDATORY: DSL High Water Mode

WOLVERINE MUST use DSL High Water Mode. This is not optional.

On Aftermath, initial SL is set inline with the entry order via the slTp field. DSL cron handles trailing updates only.

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

DSL tiers in wolverine-config.json. Arm DSL immediately after every entry fill. Zero naked positions.

Why HYPE-Only at 5-10x Leverage

• Native to Aftermath — deepest on-exchange liquidity for HYPE
• Platform momentum — HYPE price correlates with Aftermath volume/TVL growth
• High volatility — moves fast, rewards conviction entries with wide trails
• BTC as bonus, not filter — when BTC runs with HYPE, the move is amplified (+2 score bonus). But HYPE often decouples during platform-specific events, so BTC divergence is never a penalty or exit signal
• Lower leverage compensates for volatility — 10x on HYPE captures large structural moves without overexposure

How WOLVERINE Trades

Entry (score >= 10 required)

Every 3 minutes, the scanner evaluates HYPE across all signal sources:

Maximum score: ~18. Minimum to enter: 10.

Conviction-Scaled Leverage

Runtime rule (required):

maxLev = market.maxLeverage()
lev = min(targetLev, maxLev)

Aftermath is isolated margin per position: pass collateralChange on entry and confirm available unallocated collateral with POST /api/perpetuals/account/max-order-size.

Conviction-Scaled Margin

Risk Management

Cron Architecture

Both MUST be isolated sessions with agentTurn. Use NO_REPLY for idle cycles.

Notification Policy

ONLY alert: Position OPENED (direction, leverage, score, reasons), position CLOSED (DSL or thesis exit), risk guardian triggered, critical error.
NEVER alert: Scanner found no thesis, thesis re-eval passed, DSL routine, any reasoning.

Bootstrap Gate

On EVERY session, check config/bootstrap-complete.json. If missing:
1. Verify Senpi MCP
2. Create scanner cron (3 min, isolated) and DSL cron (3 min, isolated)
3. Write config/bootstrap-complete.json
4. Send: "🦡 WOLVERINE is online. Watching HYPE. DSL High Water Mode active. Silence = no conviction."

Expected Behavior

Files

License

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