←
Firepan2,256 posts
Firepan avatar

Firepan

@FirepanHQ

AI-native security for Web3 and beyond. Smart contract scans, audits, token risk intelligence, pentesting & continuous security. 🔥🍳

◷ Joined April 2024▧ 463 Photos and videos
Firepan avatar
Firepan@FirepanHQ·11h

Lighter's contracts held. A trader's homemade signer reused one nonce and gave away the private key. $10M gone. The code that drains you is rarely the code you audited. What's running in your bot that nobody has audited?

Firepan avatar
Firepan@FirepanHQ·Nov 17

This is why you need Firepan.com Fight the AI attacks with AI defense. Your first scan is free!

BREAKING: FBI warns cybercriminals & state-aligned hackers are using an AI-powered tool to automate attacks against organizations worldwide.

Ian Kane avatar
⟳ Firepan retweeted
Ian Kane@IamIanKane·Nov 17

A bank will spend 18 months on vendor risk for a tokenization pilot. Then ship smart contracts that get audited once. If the money is live 24/7, the security has to be too.

Firepan avatar
Firepan@FirepanHQ·Nov 17

Crypto teams: when did someone last try to break your web app on purpose? Not your contracts. The app. Wallet-connect flows, API keys, admin dashboards, cloud consoles. Honest answers only.

Post attachment
Firepan avatar
Firepan@FirepanHQ·Nov 17

🚨 ALERT: in the last 24 hours: • Major banks (Chase, BofA, Wells Fargo, Zelle) down from a software bug • Optimizely supply-chain attack served malware on 178+ sites (Zoom, BBC, CNN, HSBC…) • United outage delayed flights • Google confirmed the first fully automated AI patch-to-exploit pipeline AI agents are now running attacks end-to-end. Continuous defense is no longer optional.

Firepan avatar
Firepan@FirepanHQ·Nov 17

Even @Google knows, the future will require always on security auditing and pentesting.

Hackers are now using LLMs to continuously scan software updates, spot newly fixed bugs, and build attacks before teams can apply the fix — all without human involvement. That’s according to the latest report from the Google Threat Intelligence Group. We're helping defenders counter automated threats with Google AI Threat Defense, which combines Gemini and other frontier models to find and fix flaws before attackers can exploit them.

Firepan avatar
Firepan@FirepanHQ·Nov 14

If the FAA can go dark from a system issue, your protocol can go to zero from one missed vulnerability. They get “we’re investigating.” You get drained. Firepan provides robust AI code audits + full pentests (contracts, web, APIs, infra). Free scan now at Firepan.com

0:15
Firepan avatar
Firepan@FirepanHQ·Nov 12

Shipping fast was never the risk. Shipping diffs nobody reviewed is. Put review in the merge path and you get to keep the speed.

Post attachment
Firepan avatar
Firepan@FirepanHQ·Nov 11

Attackers are now targeting critical wallet repos with multi-stage malicious PRs and they’re adapting to AI reviewers. A brand-new GitHub account spent hours hitting Keplr/Vizor and Zcash (Zingo) projects: CI “diagnostics” that exfiltrated signing keys, an AGENTS.md change instructing AI tools that uploading secrets is normal, then a crates(.io) dependency with a credential stealer in build.rs (runs on cargo build/check and even rust-analyzer). This is why continuous, event-driven repo monitoring matters. Firepan scans connected GitHub workflows on PRs and dependency changes so teams catch suspicious external contributions, build scripts, and AI-instruction files before they land. Review every external PR like production code - especially AGENTS.md and build hooks.

Attackers are now publicly attacking critical repos with malicious PRs PRs. Make sure you review all external contributions comprehensively. A GitHub account created at 01:26 UTC today spent six hours trying to get into wallet repos at chainapsis (Keplr/Vizor) and Zingo Labs (Zcash). When Codex review flagged its "diagnostics" PRs that POSTed signing keys, it opened an AGENTS(.)md PR telling AI reviewers that uploading secrets is the project convention. Then it went quieter: a crate it had just published, lwd-uri-util, with the stealer in build(.)rs, slipped into Zingo PRs inside a full-file CRLF rewrite. build(.)rs runs on cargo build, cargo check, even rust-analyzer opening the project. It grabs env vars, git/gh/AWS/npm/cargo creds, Keplr MCP vaults and macOS Keychain entries. Your AI reviewer reads AGENTS(.)md too. Review it like CI config, not docs.