Josselin Feist
@Montyly
Followers
3K
Following
516
Statuses
963
Engineering director at @trailofbits. Working on blockchain security & program analysis
Joined January 2010
Join us to learn about advanced smart contract fuzzing techniques
Our most extensive fuzzing suite to date pushes the boundaries of DeFi protocol testing. Join us next Wednesday 10:30am ET for a stream as we break down the 100 invariants we developed for @Uniswap v4's test infrastructure.
0
3
11
RT @trailofbits: We have pioneered smart contract fuzzing for almost a decade. It is now a technique used by many security researchers. Wha…
0
4
0
Invariant-driven development is the future of smart contract development. This is how the industry can go to its next step of maturity and raise security significantly
Invariant-driven development is a paradigm shift: you need to integrate security properties into your software development lifecycle to catch critical vulnerabilities before they even exist.
1
5
23
Shoutout to @DeFi_Wonderland. It is really cool to see developers integrating fuzzing within their workflow and building tooling around it
We (@DeFi_Wonderland) have build a few convenience tools around Medusa (nothing fancy, we use it internally but happy to share): - youdusa to generate forge reproducers ( - Medusa-gen to generate the default template we use ( - handler for long running campaign (using discord for notif) (in cleanup phase rn, soon public;) They’re similar to others with some slight differences (ie cloud exec spins a s3 while we’re metal-based, fuzz-utils generate tests for whole corpus instead of just failing tests, etc) We’ll soon release our fuzzing campaigns and formal verification chapter in our handbook (already unit and integration tests covered:
2
8
36
RT @trailofblocks: We are starting the year strongly! Slither 0.11.0 is out! It's our most comprehensive update yet! (Find out more 👇) http…
0
18
0
RT @pcaversaccio: As a public warning, please don't interact with BeraSwap (beraswap[.]co, previously beraswap[.]fi). They rugged yesterday…
0
136
0
@milotruck @nisedo_ We definitely have a lot of room for documentation improvements. This might help: - Advanced usages at @Web3SecSummit - Hackathon we ran at @CryptoCanal's EthDam: - More exercises on the API
0
2
5
RT @milotruck: Turns out the awesome folks at @trailofbits have built slither-read-storage for this: Specifying -…
0
9
0
RT @summit_defi: Watch the "Demystifying Smart Contract Security: Facts & Fallacies" LIVE NOW: With: • Hari (@_hrk…
0
7
0
@milotruck There is a dedicated tool that comes with slither and checks for variable collisions:
0
0
4
RT @summit_defi: Dive into this panel about "Demystifying Smart Contract Security: Facts & Fallacies." during the next DSS Monthly webinar.…
0
21
0
@milotruck You might want to check out slither-read-storage ;) If you have an archive node, you can use --block BLOCK to see the value at the given block
0
0
6
@devtooligan @offbeatblog_eth I am not sure what you are referring to. We provide a lot of architectural recommendations through code maturity, appendixes, informational issues, design reviews all the time. It's true we do not rewrite the code for the client, but they don't hire us for that ;)
1
0
5
@usmannk @_hrkrshnn @NotDeGhost Can you expand on that? I would assume any guarantees provided by JUMPDEST for a jump table can be done without it. You need a few more opcodes to safeguard the jump, but you can be efficient there (ex: using bitmasking etc)
0
0
2
@_hrkrshnn @NotDeGhost Unless I am mistaken, the jumpdest analysis is not the part preventing jumping into push data. Instead, it's the fact that EVM does a pass from the first instruction and iterates over every instruction's size to know what is valid. Otherwise, you could jump into a Push 0x5b ;)
1
0
0
RT @_SEAL_Org: Happy New Year, everyone! 🎉 From securing $75M in assets to launching a legal safe harbor for white hats in 2024, our commu…
0
17
0
