Twitxy Saved

Isaac Evans

Verified account This account is verified. Learn more

@0xine

Co-founder @semgrep. Make it cheap to make it expensive to exploit software.
San Francisco, CA Joined January 2015
374 Following 710 Followers 223 Posts

Latest posts 5

Isaac Evans Verified account @0xine
Huge day for Semgrep: We’re launching Semgrep’s Agentic AppSec Platform, with Mandoline, our next-generation program analysis engine purpose-built for agents.

1/4: Mandoline: We have completely re-imagined the technology behind Semgrep. How do we build a program analysis tool that makes agents more effective? We've created Mandoline, a new engine that is 50x faster than Semgrep Pro, covers 50+ languages, and gets a perfect score on OWASP benchmarks (see the blog post). Over the past several months, we've been developing this with large customers who had early access to frontier models like Mythos and Astra. The impact is that those customers are telling us they are finding more (+40%) vulnerabilities for fewer tokens (up to ~10x reduction), with frontier labs' latest models. blog post by @drewdennison: semgrep.dev/blog/2026/semg…
5 replies 7 reposts 23 likes 4.9K views View on X
Isaac Evans Verified account @0xine
Last week I guest lectured at Stanford for Mihail Eric's class The Modern Software Developer

Was a ton of fun; a few highlights:

1/3) LLMs are the next generation of static analysis. The opportunity is to embrace this and build new tools that LLMs can leverage for faster answers and higher recall. Semgrep has a big announcement coming here!
Isaac Evans Verified account @0xine
The OSS ecosystem fundamentally changing and the frontier labs are actively talking about rewriting their *entire* dependency basis. Unthinkable a year ago and I really don't like the implications for community, especially for new & smaller OSS projects trying to get off the ground.
AI agents are writing code faster than ever

But who's securing it?

@robby_mtf & @tnachen sat down with @Semgrep co-founder @0xine to discuss why security could be AI's biggest bottleneck, the future of AppSec, and what happens when security tools are built for agents - not humans

🎙️👇
View quoted post on X
Isaac Evans Verified account @0xine
@semgrep team has been cooking! We've updated coding agent plugin (Guardian) with hooks for php2shell, LiteLLM comprompise, Node.js brotli vuln , etc. that fire automatically in agents like Claude, Codex, etc. to prevent new code & infra-code from using the vulnerability. Nice example of @claudeai using the information below to reason about conditional reachability.
Isaac Evans Verified account @0xine
If Claude Code is the new us-east-1 (I think it is), your coding agents should know about php2shell before your cloud scanner.
@Semgrep’s Guardian plugin for Claude Code, Codex & more now deterministically flags vulnerable PHP versions. Nice work team!