With an increasing number of discussions around the BIP54 “Consensus Cleanup” soft fork proposal, I helped put together an information site about BIP54.
“Bitcoin has four known vulnerabilities that have gone unfixed for 15 years. BIP54, "Consensus Cleanup", proposes four narrowly-scoped changes to address these issues in Bitcoin's consensus rules that date back to the original version of Bitcoin in 2009.”
This turned out great. It's nice to have all the discussion history available too.
Nice summaries of the problems bip54 fixes:
Stay dirty pony toy
The best part about BIP54 is the scope discipline. Four bugs, four fixes, nothing extra. No feature creep, no "while we're at it" additions. That's how you get consensus for a soft fork in a world where every proposed change gets treated like an existential threat.
The timewarp fix especially is overdue. It's the kind of vulnerability that nobody exploits because the game theory doesn't reward it right now, but "nobody would bother" is not a security model. You fix the door before someone decides to kick it in.
The site itself is well done too. Making consensus-level changes legible to non-devs is how you actually get community alignment. Most people who have strong opinions about soft forks have never read the actual BIP. This kind of resource closes that gap.
The timewarp fix is the most consequential part of BIP54. The attack: miners can set the timestamp of the last block in a 2016-block difficulty period up to 2 hours ahead, then set the first block of the next period 2 hours behind — creating a false difficulty calculation that allows mining the next 2016 blocks in much less than 2 weeks.
This was first formally documented by Friedenbach in 2018. The attack could theoretically allow a 51% miner to halve difficulty repeatedly until blocks come in seconds. BIP54 closes it with a simple timestamp constraint.