Harsh Methwani (@KernelHarsh) profile picture

AI Security @QuillAudits_AI, Red Team Engg. @GraySwanAI , prev Engg. @OverlayyAI , alma mater @IITKanpur

Earth
Joined July 2017
Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)Photo by Harsh Methwani (@KernelHarsh)
Harsh Methwani retweeted
Suicides in IITs continue. This time at IIT Roorkee. A student in second year of BS-MS course at IIT Roorkee, died by suicide on October 13, 2026.
TW: Suicide: A final-year B.Tech student at IIT (ISM) Dhanbad died by suicide on October 31, 2026. At least 53 students died --45 suicides and 8 unnatural deaths --across IITs between 2020-21 and 2024-25 hindustantimes.com/india-newโ€ฆ
2
71
6
212
19,285
Harsh Methwani retweeted
wait it all makes sense now
82
262
24
4,563
170,649
Harsh Methwani retweeted
Really liked @alexfilippov314 approach to private audits vs bug bounties. He separates them clearly: Bug bounties โ†’ ~30% manual / 70% AI Private audits โ†’ ~60% manual / 40% AI This is exactly how it needs to be now that AI is in the picture. You canโ€™t just dump the entire codebase into AI and expect it to find the real bugs. At the same time, you canโ€™t sit for hours reading every line manually youโ€™ll waste a ton of time and mostly end up finding duplicates. Creating the right balance is critical. Another point worth mentioning about Alex is the tool he built after a month of dealing with duplicates. It automates a large part of his workflow. Itโ€™s equally important to understand AI deeply enough to build these kinds of automations so that most of the repetitive tasks are handled by AI. And finally being really strong at the fundamentals is what helps you stand out. Thereโ€™s a massive difference between throwing generic prompts and giving highly specific, well contextualized ones. With the right prompts rooted in deep understanding, your chances of finding unique bugs with AI go up significantly. Good article by @b612security !
4
3
45
1,358
Harsh Methwani retweeted
We are hiring at @EFDevcon. Booth is set up at the Career Fair today, Roles in security, BD and marketing. Interns, juniors and mid level folks all welcome. Building something onchain? Come by and grab QuillShield access for your team too.
Meet the companies hiring in India and joining our Career Fair at @EFDevcon ๐Ÿ‡ฎ๐Ÿ‡ณ ๐Ÿฆ @AMINABankGlobal ๐ŸŒŠ @AquaZero0 ๐Ÿ” @dapplooker / @HyprEarn ๐Ÿš€ @DreamStarterXYZ ๐ŸŒ @libp2p ๐Ÿ›ก๏ธ @QuillAudits_AI Join the career fair on Nov 6 ๐Ÿ‘‡ dev3pack.xyz/career-fair
9
10
2
42
3,175
Harsh Methwani retweeted
I have a request for the founders, VCs and everyone attending who is in a position to help. Please make time for the developers. 8 Devcons in 12 years, and this is the first one in India. Many are travelling from other Indian cities and from abroad, for their first Devcon. Walking up to a stranger to talk about what youโ€™re building takes courage. Give them enough time to be heard. Ask about their work. Listen to where theyโ€™re stuck. Offer feedback, an intro, or help with their next step. Please donโ€™t let immediate ROI decide who gets your time. With the market mood right now, Iโ€™ve decided to spend more of my Devcon with young developers, without any ROI. Iโ€™m asking others to do the same. If youโ€™re a first-time developer there, come say hello. Iโ€™d love to hear what youโ€™re building. Iโ€™ll be in Mumbai, November 1โ€“6. Letโ€™s help them leave Mumbai with more confidence than they came with.
7
5
34
929
Harsh Methwani retweeted
quillshield just crossed 100 users in beta. it's our multi-agent ai auditor, running on the best cyber models, including claude mythos 5.1. recently it found a working exploit path in a live defi lending contract that a human auditor had ruled out: teapawt.org/QuillAudits_AI/status/โ€ฆ opening 5 free slots reply "shield" and i'll dm you a code.
QuillShield, our multi-agent AI auditor, independently discovered and wrote a working exploit for a bug in a live lending contract, after a 12-agent AI pass had found the bug and a human auditor had stopped at the safeguard. The auditor later rebuilt the exploit on a mainnet fork and proved it works. The auditor is @Harvesto12, a security researcher. His 12-agent pass had already flagged the bug. He tested one attack path, a compensating safeguard elsewhere in the system blocked it, so he filed the bug as refuted and moved on. QuillShield found the same bug on its own, but it came at it from a different path, one the safeguard did not cover, and then wrote its own exploit test for it. Then he took that exploit test and rebuilt it himself on a Foundry mainnet fork, with a fresh account and no injected state, so nothing was staged. It went through. The rest of the run, in his numbers โ€ข 8 findings, 4 of them his pass had not caught โ€ข 3 of the 8 exploit tests it wrote and ran passed โ€ข severity on the bug went Medium from QuillShield, Critical from him 4 reviewers, and not one of them disputed that the bug is real. His explanation for why it got through where he hadn't is the part I keep thinking about. He had tested one path and assumed the safeguard handled the rest. QuillShield had no such assumption, so it worked through the other paths until one went past it. That's the point of building QuillShield: to keep challenging the code when one blocked attack makes it look safe.
14
6
3
27
1,602
Harsh Methwani retweeted
QuillShield, our multi-agent AI auditor, independently discovered and wrote a working exploit for a bug in a live lending contract, after a 12-agent AI pass had found the bug and a human auditor had stopped at the safeguard. The auditor later rebuilt the exploit on a mainnet fork and proved it works. The auditor is @Harvesto12, a security researcher. His 12-agent pass had already flagged the bug. He tested one attack path, a compensating safeguard elsewhere in the system blocked it, so he filed the bug as refuted and moved on. QuillShield found the same bug on its own, but it came at it from a different path, one the safeguard did not cover, and then wrote its own exploit test for it. Then he took that exploit test and rebuilt it himself on a Foundry mainnet fork, with a fresh account and no injected state, so nothing was staged. It went through. The rest of the run, in his numbers โ€ข 8 findings, 4 of them his pass had not caught โ€ข 3 of the 8 exploit tests it wrote and ran passed โ€ข severity on the bug went Medium from QuillShield, Critical from him 4 reviewers, and not one of them disputed that the bug is real. His explanation for why it got through where he hadn't is the part I keep thinking about. He had tested one path and assumed the safeguard handled the rest. QuillShield had no such assumption, so it worked through the other paths until one went past it. That's the point of building QuillShield: to keep challenging the code when one blocked attack makes it look safe.
5
8
6
35
3,175
Beta access has been opened to QuillShield, a multi-agent auditor for EVM smart contracts, check it out.
Scan your smart contracts before attackers do. We are opening beta access to QuillShield, our multi-agent auditor for EVM smart contracts, for the first 50 teams. October has made it clear that waiting is no longer an option. In the last few days alone: โ€ข PHP: pre-auth remote code execution, CVSS 9.2 โ€ข GitLab: CVSS 10, self-hosted servers back to 2014 โ€ข Node.js: 10 CVEs in one release โ€ข Railway: customer secrets accessed โ€ข LiteLLM: credential stealer shipped on PyPI AI helped find some of these bugs, and attackers have the same tools. A bug in a deployed smart contract can drain funds in a single transaction, so we built QuillShield to find them first. Here's what it has done so far: > 25 Critical or High vulnerabilities caught across 155 projects > 497 findings, with 85% of auditor-reviewed ones confirmed valid > 350,000+ lines of Solidity reviewed > 83.8% on EVMBench, against 45.6% for the best published model Here's the beta link: shield.quillaudits.com/
3
2
18
2,827
Harsh Methwani retweeted
Scan your smart contracts before attackers do. We are opening beta access to QuillShield, our multi-agent auditor for EVM smart contracts, for the first 50 teams. October has made it clear that waiting is no longer an option. In the last few days alone: โ€ข PHP: pre-auth remote code execution, CVSS 9.2 โ€ข GitLab: CVSS 10, self-hosted servers back to 2014 โ€ข Node.js: 10 CVEs in one release โ€ข Railway: customer secrets accessed โ€ข LiteLLM: credential stealer shipped on PyPI AI helped find some of these bugs, and attackers have the same tools. A bug in a deployed smart contract can drain funds in a single transaction, so we built QuillShield to find them first. Here's what it has done so far: > 25 Critical or High vulnerabilities caught across 155 projects > 497 findings, with 85% of auditor-reviewed ones confirmed valid > 350,000+ lines of Solidity reviewed > 83.8% on EVMBench, against 45.6% for the best published model Here's the beta link: shield.quillaudits.com/
20
19
25
77
16,147
Harsh Methwani retweeted
Seems like we've been noticed too!! Not made any updates yet but we are cooking something. Soonnnn....
We've been listed on ethereum.org for our contributions to the Ethereum ecosystem with our AI agent Skills repository This is just the beginning of our contributing, planning so much more, looking to grow the pie for everyone around๐Ÿซก
4
28
858
Harsh Methwani retweeted
QuillAudits now has @AnthropicAI Defense Access, under its Cyber Verification Program. Most QuillShield agents and our audit team now work with it. Attackers are already using AI on onchain code. QuillShield is the defensive side, now with verified access. QuillShield is our AI security reviewer for onchain finance. It reviews every pull request with the full protocol in context, the way Greptile does for general codebases. Beta is open to selected teams, Solidity first. DM us for access.
5
7
1
34
1,427
Harsh Methwani retweeted
๐Ÿคฏ ~24k smart contract audit findings just got open-sourced Titles, bug descriptions, PoCs, recommendations + severity ratings from private audits & contests A massive dataset for auditors, researchers and AI security builders Go dig in by @RightNowIn huggingface.co/datasets/Zaevโ€ฆ
13
47
3
314
22,544
Harsh Methwani retweeted
. @KernelHarsh already cashed a Gray Swan bounty and is now in Hazard Hunt โ€” $80k to stress-test Chem, Bio, and Cyber refusals on frontier models. Findings go straight back to the labs. Which category do you think breaks first?
1
12
419
Harsh Methwani retweeted
12 projects. 155 findings. 0 missed deadlines. That was August at QuillAudits ๐Ÿš€ 5 critical 24 high 42 medium 50 low 34 informational All 155 reported to the team, all 155 resolved before sign-off. Shipping in Q4? No audit booked yet. Talk to us soon . September is already filling up ๐Ÿ‘‡ calendly.com/akshadh-a-quillโ€ฆ
14
6
2
38
3,047
Harsh Methwani retweeted
Stop jailbreaking cloud AIs and risking account suspension. Start running uncensored AI on your laptop or even on your phone.
Stop using English to jailbreak LLMs. Stop using English to jailbreak LLMs. Stop using English to jailbreak LLMs. Stop using English to jailbreak LLMs. Stop using English to jailbreak LLMs. Stop using English to jailbreak LLMs. Stop using English to jailbreak LLMs.
16
5
200
16,066
๐Ÿ“‚ Useful Crypto Apps (2026 edition) โ”ƒ โ”ฃ ๐Ÿ“‚ DeFi โ”ƒ โ”ฃ ๐Ÿ“ @aave โ”ƒ โ”ฃ ๐Ÿ“ @MorphoLabs โ”ƒ โ”ฃ ๐Ÿ“ @compoundfinance โ”ƒ โ”ฃ ๐Ÿ“ @LidoFinance โ”ƒ โ”ฃ ๐Ÿ“ @kamino โ”ƒ โ”— ๐Ÿ“ @jito โ”ƒ โ”ฃ ๐Ÿ“‚ Wallets โ”ƒ โ”ฃ ๐Ÿ“ @MetaMask โ”ƒ โ”ฃ ๐Ÿ“ @phantom โ”ƒ โ”ฃ ๐Ÿ“ @Rabby_io โ”ƒ โ”ฃ ๐Ÿ“ @rainbowdotme โ”ƒ โ”ฃ ๐Ÿ“ @DinhtienSol โ”ƒ โ”ฃ ๐Ÿ“ @TrustWallet โ”ƒ โ”— ๐Ÿ“ @safe (multisig) โ”ƒ โ”ฃ ๐Ÿ“‚ Hardware Wallets โ”ƒ โ”ฃ ๐Ÿ“ @Ledger โ”ƒ โ”— ๐Ÿ“ @Trezor โ”ƒ โ”ฃ ๐Ÿ“‚ Stablecoin Yield โ”ƒ โ”ฃ ๐Ÿ“ @pendle_fi โ”ƒ โ”ฃ ๐Ÿ“ @ethena โ”ƒ โ”— ๐Ÿ“ @sparkfinance โ”ƒ โ”ฃ ๐Ÿ“‚ Onchain Perps โ”ƒ โ”ฃ ๐Ÿ“ @HyperliquidX โ”ƒ โ”ฃ ๐Ÿ“ @Lighter_xyz โ”ƒ โ”ฃ ๐Ÿ“ @Aster_DEX โ”ƒ โ”— ๐Ÿ“ @GMX_IO โ”ƒ โ”ฃ ๐Ÿ“‚ Swaps & Aggregators โ”ƒ โ”ฃ ๐Ÿ“ @Uniswap โ”ƒ โ”ฃ ๐Ÿ“ @JupiterExchange โ”ƒ โ”ฃ ๐Ÿ“ @1inch โ”ƒ โ”— ๐Ÿ“ @CoWSwap โ”ƒ โ”ฃ ๐Ÿ“‚ Bridges โ”ƒ โ”ฃ ๐Ÿ“ @AcrossProtocol โ”ƒ โ”ฃ ๐Ÿ“ @StargateFinance โ”ƒ โ”ฃ ๐Ÿ“ @lifiprotocol โ”ƒ โ”ฃ ๐Ÿ“ @RelayProtocol โ”ƒ โ”— ๐Ÿ“ @deBridgeFinance โ”ƒ โ”ฃ ๐Ÿ“‚ Crypto Cards โ”ƒ โ”ฃ ๐Ÿ“ @ether_fi โ”ƒ โ”ฃ ๐Ÿ“ @MetaMask โ”ƒ โ”ฃ ๐Ÿ“ @gnosispay โ”ƒ โ”ฃ ๐Ÿ“ @RedotPay โ”ƒ โ”ฃ ๐Ÿ“ @cryptocom โ”ƒ โ”ฃ ๐Ÿ“ @Plasma โ”ƒ โ”ฃ ๐Ÿ“ @KASTxyz โ”ƒ โ”— ๐Ÿ“ @holyheld โ”ƒ โ”ฃ ๐Ÿ“‚ Tokenized Stocks โ”ƒ โ”ฃ ๐Ÿ“ @RobinhoodApp โ”ƒ โ”ฃ ๐Ÿ“ @xStocksFi โ”ƒ โ”ฃ ๐Ÿ“ @Ondo โ”ƒ โ”— ๐Ÿ“ @Gemini โ”ƒ โ”— ๐Ÿ“‚ Prediction Markets โ”ฃ ๐Ÿ“ @Polymarket โ”ฃ ๐Ÿ“ @Kalshi โ”— ๐Ÿ“ @trylimitless Part 2 ๐Ÿ‘‡
121
129
26
918
100,694
Harsh Methwani retweeted
Every AI security tool claims it finds criticals. Very few put that claim in a public arena where findings are judged, deduplicated, and scored next to human researchers on the same code. That is why QuillShield (our multi-agent Security Auditor) is participating across the Audit Engine alongside @Certora, @v12sec and others. If your AI auditor has never been judged on someone else terms, you do not know what you have.
As more Audit Engine engagements begin opening up, we want to share a little more about how they actually work from a participant perspective. A quick thread on the formats, timelines, rewards, judging, and how to get involved โ†“
2
6
31
1,414
Harsh Methwani retweeted
Spent last week triaging and improving. Was worth it! ๐Ÿ†
1
3
172
Harsh Methwani retweeted
Modi ji, Remove Dharmendra Pradhan tomorrow. Thatโ€™s the strictest action you can take.
845
1,556
85
7,613
163,387