The Circuitry
THE CIRCUITRYYour one-stop source for all tech news
  • Today
  • Feed
  • Events
  • Saved
HOMETODAYFEEDEVENTS
BOOKMARKS
RSS
© 2026 The Circuitry
About UsSourcesContactCorrectionsPrivacy
  • Today
  • Feed
  • Events
  • Saved
  • AAPL
  • MSFT
  • GOOGL
  • META
  • NVDA
  • TSLA
  • AMZN
  • INTC
  • SONY
  • TTWO
  • COIN
  • BTC
  • ETH
  • SOL
  • AAPL
  • MSFT
  • GOOGL
  • META
  • NVDA
  • TSLA
  • AMZN
  • INTC
  • SONY
  • TTWO
  • COIN
  • BTC
  • ETH
  • SOL

Feed

Ad
From The CircuitryVerified tech news, cross-checked.How we verify →
  • Google links COLD RELIC's 2025 US breach to a second intrusionTech
  • NVIDIA's third-quarter revenue rises 103% to $115.8 billionTech
  • Google launches a Robotics plugin for Gemini 4 HydrogenTech
  • Google Wallet is now available on iOS in the U.S.Tech
  • Two Grok Bots made 31 unattended requests for employee dataTech
  • Kinsus plans more than 3.5x substrate capacity for NVIDIATech
  • Microsoft's Ignite keynote adds Copilot Workspaces and Cobalt 200Tech
  • Apple says an internal update caused today's service outagesTech
  • Jensen Huang says he would return as Nvidia CEO if directors decideTech
Scroll for more
Verification
VERIFIEDConfidence: HIGH
Source identified
Claims cross-referenced
No discrepancies found
Sourcing
1source
Google · track record
16Stories
100%Verified
1430d
All sources →
Markets
GOOGL···

Live quote · not investment advice

Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
Home/Tech/Google finds an AI agent that built exploits with no human in the loop
VERIFIEDBy Xavier Rivera·Monday, November 16, 2026 at 8:22 PM AST·2 min read

Google finds an AI agent that built exploits with no human in the loop

Google has identified a framework that drafted and ran exploits with no person in the loop. The same report says zero-day exploits rose to 22 in August, and stolen model access reached $1 million in the worst cases.

Source:Google
Post
Google finds an AI agent that built exploits with no human in the loop
Ad
From The CircuitryVerified tech news, cross-checked.How we verify →
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
TL;DRAI · 60 sec read

Google has identified a framework named exploit_agent that compared patched code, wrote exploits, and ran them with no person in the loop. The same report says CAIRN and PentAGI were used in live intrusions, and zero-day exploits rose to 22 in August. Stolen model access passed $100,000 in a day in some cases and reached $1 million in the worst.

Morning Brief

Get the daily Google brief

Every morning, in your inbox. What's in The Brief →

Google Threat Intelligence Group has identified a framework that drafted and ran exploits with no person in the loop. The find, published on 16 November, is the first time the group has seen threat actors fully automate vulnerability discovery and exploitation.

The framework is named exploit_agent, and it sat on an exposed staging server. It connects to several frontier and local models. It watches popular web application software, compares the code before and after a patch announcement, writes an exploit from that difference, and runs the exploit on its own.

The server held data taken from targeted victims. Scripts on it can map an internal network, try passwords, install webshells, and remove webshells left by other attackers. Telemetry on the server points to French sites in media, entertainment, and education.
The server held data taken from targeted victims.

The configuration file includes comments written in Russian. Those comments label the keys as being for automatic exploitation by the framework. Google does not name the operator.
Timeline of automated vulnerability research
Timeline of automated vulnerability research · Google

The same report says the past quarter is the first time this full automation showed up. The pipeline takes a published patch, finds what changed, writes an exploit, and runs it. That is the step the group had not seen closed without a person before.
From The CircuitryThe Feed — live briefs across tech, all day.See what’s happening →

The staging server was left open, which is how the framework was found. Google says the operator used it to stage the automated work and to hold what the exploits returned. The report describes the server, the scripts, and the stolen data. It does not name a victim company.
The configuration file includes comments written in Russian.

The same report says two open-source attack tools, CAIRN and PentAGI, have been used in live intrusions. CAIRN was used against targets in the U.S. aerospace sector. A customized PentAGI platform scanned organizations across Asia from late August through early October.

Zero-day exploits averaged 8 a month in 2025 and 11 a month in 2026, then rose to 22 in August. Overall disclosures and exploits nearly doubled from January to August. In separate cases of stolen model access, a single day's fraudulent charges passed $100,000, and the worst cases reached $1 million before they were stopped.
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
A suspected group changed Claude Code settings so that opening the tool downloaded a backdoor. The targets were cryptocurrency organizations in North America and Europe. Google says it disabled assets tied to the abuse it found.
Why this mattersAI · ~100 words

Tap a lens to see what this story means for you.

Morning Brief

Liked this? The Brief brings you the whole day in tech, verified, every morning.

Two minutes, free forever. What's in The Brief →

Reader-supported
DonateBuy me a coffee →Follow@thecircuitry_ →Follow@thecircuitry.to →
HELP US IMPROVE
Grand Theft Auto VILive hubUS store unlocks Thu Nov 19, 1:00 AM ASTFirst unlock: New Zealand · Wed 7:00 AM ASTLaunch hub →Image: Rockstar Games
From The Circuitry

See what’s happening right now

The Feed runs all day — short, verified briefs the moment they break.

Open the Feed →
From The Circuitry

Follow @thecircuitry_

Every story we publish, as it happens. No noise between.

Follow on X ↗On Bluesky ↗

Reader-supported

The Circuitry is a passion project I've always wanted to build, and I love the work behind it.

Running it costs real money. APIs, hosting, time. To keep improving the site and growing this into something useful for everyone, those costs have to be covered.

Any contribution is appreciated. If not, no pressure. Thanks for reading.

Buy me a coffee
googlesecurityai
Grand Theft Auto VILive hubUS store unlocks Thu Nov 19, 1:00 AM ASTFirst unlock: New Zealand · Wed 7:00 AM ASTLaunch hub →Image: Rockstar Games
More inTech
  • Kinsus plans more than 3.5x substrate capacity for NVIDIA

    Tech · 18m
  • Microsoft's Ignite keynote adds Copilot Workspaces and Cobalt 200

    Tech · 2h
  • Apple says an internal update caused today's service outages

    Tech · 4h
SupportThe Work

The Circuitry is reader-supported. If you find the daily brief useful, you can buy me a coffee to keep it going.

Buy a coffee →
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →

MORE IN THIS BEAT

All Tech →
  • Tech·Nov 17

    Google links COLD RELIC's 2025 US breach to a second intrusion

    Google has linked a late-2025 data theft at a US organization to a 2026 intrusion at a second US organization by the Russian group COLD RELIC. Operators used stolen staff records from the first breach to impersonate an employee and phish the next target.

  • Tech·Nov 17

    NVIDIA's third-quarter revenue rises 103% to $115.8 billion

    NVIDIA has reported third-quarter revenue of $115.8 billion, up 103% from a year ago, with data-center revenue of $107.6 billion. Fourth-quarter revenue is expected at $130.0 billion, plus or minus 3%, with no data-center compute revenue from China assumed.

  • Tech·Nov 17

    Google launches a Robotics plugin for Gemini 4 Hydrogen

    Google has launched a Robotics plugin for Antigravity that runs Gemini 4 Hydrogen on a supported arm. The plugin is limited to paid Google AI Pro and Ultra subscribers, and a demo arm was ready for tasks in about 10 minutes.

  • Tech·Nov 17

    Google Wallet is now available on iOS in the U.S.

    Google has opened Google Wallet on iOS in the United States. The app stores passes and payment cards, and tap to pay is not supported on iOS.

  • Tech·Nov 17

    Two Grok Bots made 31 unattended requests for employee data

    Two Grok Bots made 31 unattended requests for employee calendar, email, Slack, GitHub, and Notion data over five days. Every request was a read, and none of them reached the data.