whale_trade_tracking v1.0.0

name: whale_trade_tracking description: Tracks large on-chain transfers and block trades, and reads whale flow without getting faked out. Triggers when a user asks to follow whale wallets, set up large-transaction alerts, interpret exchange inflows/outflows, or understand dark pool prints and OTC/RFQ block trades.

Whale Trade Tracking

Purpose

Give an agent a repeatable workflow for finding large on-chain transfers and block trades, attaching entity context to them, and reading what the flow actually implies. Large transfers are cheap to observe and easy to misread; this skill keeps the reading honest.

Workflow

1. Define the watch. Scope first: which asset (BTC, ETH, a token, a stablecoin), which threshold counts as whale-size for that asset, and what question the read has to answer. A read without a question is noise.
2. Pick data sources for the lane. On-chain lanes: block explorers with labeled addresses (Arkham, Etherscan family), alert feeds (Whale Alert), DeFi portfolio views (DeBank, Zapper), raw SQL access (Dune, Flipside, The Graph). Block-trade lane: RFQ/OTC desks (e.g. FalconX Electronic Options, prime-broker RFQ networks), dark pool prints on the consolidated tape (crypto ETFs like IBIT print through Nasdaq), iceberg and hidden-order reading on the CEX tape. See references/data-sources.md for the full map.
3. Attach entity context before interpreting. A raw address is a guess. Check whether the sender or receiver is labeled (exchange hot/cold wallet, OTC desk, fund, bridge, protocol treasury) on Arkham, Etherscan, or DeBank. Follow the receiving address's history: other wallets owned by the same entity often surface in inflow/outflow tabs. If no entity is identifiable, the read is informational only, not a trading signal.
4. Read the flow, not just the size. Exchange-flow reading: large deposit to an exchange means someone is positioned to sell (bearish setup); large withdrawal from an exchange means coins leaving the sellable pool (bullish setup); stablecoin deposits to an exchange mean someone is positioned to buy (bullish setup); stablecoin withdrawals mean stables moving to DeFi or cold storage (neutral to bullish); a big transfer between two unlabeled wallets is often internal housekeeping and means nothing. See references/flow-reading.md for the full table.
5. Read block prints and hidden size. Block prints: negotiated trades (OTC desks, dark pools) are executed in secret and then reported to the public tape, sometimes with delayed reporting. When a print lands, market makers adjust delta hedges in the lit market, which can push price in the print's direction short term. Direction matters more than size: a block bought above the ask reads as institutional buying; one sold below the bid reads as selling. Iceberg tells on the tape: a single big-volume bar with little price impact means size was absorbed; repeated equal-size prints or a bid/ask that refills at the same price after fills means hidden size; a large order that pulls when price approaches is spoofing, not support.
6. Build the alert pipeline. Set entity-level alerts (Arkham custom alerts, or a free Whale Alert feed) for the wallets and thresholds from step 1; route them to Telegram, Slack, or email. For custom automation, poll a block explorer API on a timer and page on threshold crossings. Keep thresholds tight enough that each alert is worth opening; alert fatigue kills the whole pipeline. See bin/whale_scan.py for a working polling example.
7. Deliver the read in the Output Contract. Never say "whale bought X so price goes up." Say who moved what, in which direction, with what entity context, and what that setup historically implies. Name every unknown explicitly.

Output Contract

A whale-activity read must contain:

• The move: asset, amount, USD estimate at time of transfer, sender, receiver, tx hash or block print reference.
• Entity context: labeled owner on both sides, or a plain statement that the entity is unknown.
• The setup: one of the flow interpretations (accumulation, distribution, buy-pressure setup, risk-off, internal/noise), or "unknown."
• Confidence and unknowns: what is verified on-chain vs. what is inferred, and what would change the read (e.g. "if the receiver is an exchange hot wallet, this is sell positioning; unlabeled, it is noise").
• Sources: explorer or alert-feed links for each claim.

Operating Rules

1. Size without entity context is noise. Large operational transfers happen constantly; never treat a transfer as directional until the counterparty is labeled or the pattern repeats.
2. Never claim a whale's motive from one transfer. Accumulation and distribution are sequences, not single events. One deposit is a setup, not a verdict.
3. Treat unlabeled-wallet transfers as informational. Do not build signals on them.
4. Internal transfers, exchange wallet shuffles, and bridge moves look like whale activity and mean nothing. Check the receiver's history before sounding the alarm.
5. Block prints need direction. Size alone tells you institutions were active; whether the block was bought above the ask or sold below the bid tells you what they wanted.
6. This is market intelligence, not financial advice. Say that when the output could be acted on. Do not invent prices, dates, quotes, or entity attributions you did not verify.