Light

Episode #1088 Quiz

A nefarious novel use for AI
Date: 2026-07-21 | Length: 2.75 hrs | Episode page at twit.tv

About this episode

Microsoft’s July Patch Tuesday fixed 570 flaws, including 145 remote-code-execution and 254 privilege-escalation bugs, plus three 0-days. OpenSSL’s HollowByte DoS lets unauthenticated attackers send 11 TLS bytes to trigger oversized preallocation, heap fragmentation, and memory exhaustion until process restart. WordPress forced updates for wp2shell pre-auth RCE.

Your name and email are stored only in your browser local storage for convenience. They are not retained server-side.

NameEmail

Question 1: According to the episode, what specific change did the OpenSSL team make to mitigate HollowByte?

Question 1

According to the episode, what specific change did the OpenSSL team make to mitigate HollowByte?

They switched to incremental buffer growth based on bytes actually received They moved all TLS buffer allocation into user space retry logic They disabled TLS handshake pre-allocation for all clients They capped every handshake buffer at 11 bytes regardless of claimed size

Question 2: What did the episode say made the OpenSSL DoS especially hard to resist with ordinary connection limits?

Question 2

What did the episode say made the OpenSSL DoS especially hard to resist with ordinary connection limits?

It used long-lived authenticated sessions that bypassed rate limits It fragmented memory so badly that RSS kept climbing even after disconnects It exhausted CPU by forcing repeated certificate validation It caused kernel panics by corrupting the network stack buffers

Question 3: In the 1Password-for-Claude flow described in the episode, what happens after the user approves the login request?

Question 3

In the 1Password-for-Claude flow described in the episode, what happens after the user approves the login request?

Claude receives the vault item and stores it in its context 1Password injects the credential directly into the page without exposing it to Claude The user manually types the password after biometric approval The password is copied into the browser extension and then masked from Claude

Question 4: What did Bitwarden identify as the key security problem with AI agents and credentials?

Question 4

What did Bitwarden identify as the key security problem with AI agents and credentials?

Agents are unable to complete any browser task unless secrets are reissued permanently Agents can only operate safely if MFA is removed from all workflows Agents cannot use passwords unless they are stored in plaintext locally Credentials may need to visit the AI agent as plaintext during use

Question 5: What was the novel AI use case the episode highlighted for FulcrumSec and similar ransomware groups?

Question 5

What was the novel AI use case the episode highlighted for FulcrumSec and similar ransomware groups?

Using LLMs to analyze stolen data and to negotiate more professionally with victims Using LLMs to identify exploitable memory corruption bugs before encryption Using LLMs to generate phishing lures for initial access campaigns Using LLMs to automate ransomware encryption so victims cannot restore files

Cancel

My Data//Security & Privacy//Clear Local Data

© 2026 Ken ArmstrongSecurity Now™ is a trademark of TWiT LLC. All trademarks and copyrights of the Security Now podcast are respected and reserved to their respective holders.
This is a hobby project made with love and admiration for what Steve Gibson and Leo Laporte have built over 20 years.