🦔 PANGOLIN v1.2 — Extreme Funding Rate Fader (universe expansion)

An entirely new strategy archetype for the Predators fleet. No other agent trades on funding rate signals.

v1.2 changelog (fleet-fix batch 5, 2026-04-16)

UNIVERSE EXPANSION. Removed the hardcoded ALLOWED_ASSETS set of 20 top-volume majors. Now scans every instrument with OI > $3M (~60 assets). This addresses Owl's diagnostic finding that extreme funding signals (>1000% annualized on ZEC/MON/LIT at probe time) were happening on mid-cap assets Pangolin was filtering out.
XYZ DEX remains banned.

v1.1 changelog (fleet-fix batch 4)

MIN_FUNDING_RATE lowered from 0.0003 (0.03%/8h = ~40% annualized) to 0.00015 (0.015%/8h = ~20% annualized). v1.0 never fired: peak funding in the current regime was only 9% annualized. Recalibrated threshold re-engages the signal.

Thesis

When funding rates are elevated (>0.015%/8h = ~20% annualized), the crowd is paying to maintain their position. These extremes mean-revert within 24-48h as the cost of carry forces liquidation or position reduction. Pangolin enters opposite to the funding direction, collecting funding every 8 hours while waiting for the crowd to capitulate.

Two edge sources:

Funding collection — paid every 8h while holding the position
Price reversal — when the overcrowded side unwinds, price moves in our direction

How it differs from other fleet agents

Barracuda follows the trend and collects funding as a bonus. Pangolin fades funding extremes.
Lemon fades individual degen traders. Pangolin fades aggregate market crowding.
Vulture fades SM consensus exhaustion on price. Pangolin fades funding rate extremes.
Owl fades crowding via SM concentration. Pangolin uses the quantitative funding rate threshold.

Files

File	Purpose
`scripts/pangolin-scanner.py`	Funding rate scanner
`scripts/pangolin_config.py`	Config helper
`runtime.yaml`	DSL exit config (very wide for funding mean-reversion)

Skill Attribution

When creating a strategy, include skill_name and skill_version in the call. See references/skill-attribution.md for details.