remora-strategy

Installation
SKILL.md

🐟 REMORA v1.0.0 β€” Whale Single-Position Mirror

Ride the whales you choose. A remora fish attaches to a shark and goes where it goes. Remora attaches to a small, hand-picked set of whale traders β€” it finds each whale's biggest-conviction bet and mirrors the strongest one, with a consensus boost when several whales pile into the same trade.

Why this strategy exists

The broad trader-followers (Raptor, Jackal, Spider) scan a leaderboard universe and synthesize a pick. That's powerful but opaque β€” you don't control whom you're following. Remora is the opposite: you name the whales. It mirrors their single highest-conviction position, so your exposure tracks specific traders you trust, and it leans hardest when those whales agree.

CRITICAL RULES

RULE 1: You pick the whales

Remora trades nothing until config.whales is set to a list of trader_ids / wallets. There is no default whale set β€” these are your choices.

RULE 2: Highest conviction = largest notional

For each whale, Remora takes their single largest-notional open position (size Γ— entry, falling back to margin), above minNotionalUsd (default $5,000, to skip dust). That's the whale's strongest live bet.

RULE 3: Consensus is the edge multiplier

Top positions are aggregated into (asset, direction) candidates across whales. The more whales independently hold the same asset+direction, the higher the score (2 whales β†’ +2, 3+ β†’ +3). One whale can be wrong; three agreeing is a much stronger signal.

Installs
4
GitHub Stars
99
First Seen
Jun 7, 2026
remora-strategy β€” senpi-ai/senpi-skills