Tendril exploit drains roughly $555K in USDT after sell hook flaw

加载图片中

Photo: crazy motions / Pexels

An attacker abused a bug in the TENDRIL token's sell logic to pull nearly all USDT from its PancakeSwap V2 pool on BNB Chain

Thu 5 Nov 2026 · 02:31 EST · 3 min read

In Crypto

Share

A small token on BNB Chain just learned an expensive lesson about custom code. A flaw in TENDRIL’s sell hook let an attacker withdraw nearly all of the USDT backing the token, producing a net loss of about $555,000.

The exploit hit the TENDRIL/USDT liquidity pool on PancakeSwap V2 on November 5, 2026.

How the TENDRIL attack worked

The weak point was the token’s sell logic. Many tokens on BNB Chain add custom behavior that fires whenever someone sells, such as fees, burns or reserve adjustments. That extra code is usually called a hook.

Security researchers traced the attack to function calls including PoolControl.sellToken, paired with a controlled burn operation. According to the disclosures, the attacker used that burn to manipulate the token’s reserves before external transfers had finished executing.

The contract let the attacker change the numbers the pool relied on for pricing while a sale was still in progress. Once those reserves were distorted, the pool treated TENDRIL as far more valuable relative to USDT than it should have been.

Automated market makers like PancakeSwap price assets based on the ratio of tokens sitting in a pool. Shrink one side artificially and the other side suddenly looks cheap to withdraw. The attacker used that gap to pull out nearly the entire USDT balance.

One detail stands out from the on-chain data. During the exploit, the attacker sent 30k USDT to a hardcoded beneficiary address.

The numbers do not quite agree

Loss estimates vary depending on who is counting. The headline figure is a net loss of about $555,000, with research findings citing roughly $550,000 in USDT drained in total.

The news moving money, markets, and the world—before your day starts.

Daily. Free. Join 34,000+ readers across crypto, finance, and policy.

We respect your privacy. Unsubscribe anytime.

Early alerts painted a smaller picture. Monitoring service TenArmorAlert initially pegged losses at approximately $309,000.

The gap reflects differing approaches to calculating net losses versus attacker profits. A pool’s total drained balance, the attacker’s take after costs, and the victims’ net loss can all produce different totals from the same transactions.

Security teams at ExVul and ScaleBit issued public alerts about the attack on November 5, 2026. Their disclosures included the vulnerable token contract, the affected pair address and the relevant transaction data, giving other researchers a trail to follow.

Background: custom tokens, custom risks

TENDRIL had little visibility before this incident. The pattern is familiar on BNB Chain. Developers bolt tax, burn or reflection logic onto standard token templates to create a distinctive tokenomics story. Each modification adds code paths that interact with the AMM in ways the original templates never anticipated.

When those paths let a user alter pool reserves mid-transaction, the AMM’s pricing math stops protecting liquidity providers.

CertiK continues to track incidents like this one as part of its Web3 security dashboard, which monitors losses, attack vectors and chain-level risk across quarters. The TENDRIL drain now joins that running tally for the second half of 2026.

What this means for DeFi users and builders

For TENDRIL holders, the immediate problem is liquidity. With nearly all USDT withdrawn, there is very little left on the other side of the pool to absorb sells.

For liquidity providers more broadly, a well-audited exchange cannot protect a pool from a token that rewrites its own reserves.

Builders face a clearer takeaway. Sell hooks and burn functions that touch pool balances need strict ordering, so that state changes happen only after transfers are complete. The TENDRIL exploit appears to have turned on exactly that sequencing failure, the kind of issue independent audits are designed to catch.

Disclosure: This article was edited by John Chen. For more information on how we create and review content, see our Editorial Policy.