The next era of Web3 infrastructure deserves a brand new type of security review.
For @0xPolygon's Heimdall v2 upgrade, Sherlock is bringing that model to life.
June 15 to July 6.
day 396/0 crits found
i've been continuing my journey building a coverage-guided smart contract fuzzer from scratch with revm. for learning purposes.
i published two blog posts this week about the topic
i have a bunch of ast-grep rules. planning to move them to a custom cli this weekend so i can fully control the output (reduce token usage, etc).
wasn't sure Kimi K2.6 could one-shot this, turns out it actually can ... pretty impressed bro
feels like the custom compaction is a key reason these long-running tasks work well. i borrowed the technique from codex and added some adjustments on top (skill-aware compaction, etc)
this graph perfectly shows how hyped i was for @FireworksAI_HQ fire pass v2
so far so goooood, <1s ttft & >150 tps. hit a few rate limits on the first day, needs to adjust my workflow and my harness retries need some work
aside from that, zero issues. barely touching GPT 5.5 now
i can't relate to this ๐
my harness is just simply bash edit_file bwrap-based sandbox (no internet access)
never create any plan, i just simply chat her like my coworker ๐คฃ
kimi k2.5 turbo on @FireworksAI_HQ is actually usable bro
full debug workflow on my linux sandbox library in under 5 mins
from finding root cause, bug repro, to fix in one session. crazy fast
figma: perfect monochrome palette
zed terminal: everything looks the same
spent way too long tweaking hex values
it was @zeddotdev auto-adjusting contrast the whole time ๐ซ