Coin Brief ENDE

Failures at the edges: a mining dashboard, an empty BUSD market and a lock on old blocks

None of this weekend's crypto incidents touched a consensus rule. Each came from a part of the system that sits beside the protocol and gets less attention.

At Luxor, attackers went through Commander, the company's tool for managing mining machines, and pointed customers' hashrate at the OCEAN pool. Luxor says just over a dozen accounts were affected, the tool was patched within two hours and Luxor Pool and LuxOS were not involved. The damage was done by changing where machines send their work, a setting that sits in a management dashboard rather than in Bitcoin itself. A flagged address has since received two OCEAN payouts of about 0.162 BTC in total.

At Kava Lend, the weak point was a BUSD market that was barely used. According to the on-chain analysis NewsBTC cites, pushing that market's net supply close to zero made the protocol overstate deposits about 243,000 times, which allowed 127 borrows totalling about 55.2 million KAVA. Kava says the event was confined to Kava Lend. A dormant market had stayed open long enough to become the entry point.

In go-ethereum, the issue was not an attack but a finding from a penetration test. Reads of old chain data through chainFreezer.ReadAncients took an exclusive lock, so RPC and peer requests for historical blocks queued every other reader on the node. The fix, merged on 16 November, changes one lock to a shared one; a second finding from the same test stops geth from refusing to start with a config file that contains a removed field.

Failures at the edges: a mining dashboard, an empty BUSD market and a lock on old blocks
Failures at the edges: a mining dashboard, an empty BUSD market and a lock on old blocks — Coin Brief

What to watch

The common thread is that the exposed parts were operational: a management tool, an unused market, a locking detail and a configuration file. Luxor has promised a post-mortem and Kava has not yet published one. Those reports will show whether the entry points were specific to each service or patterns other operators share.

Written by Victoria Shinder.