BTC $71,807
2026 Bull Run Is Building Start trading with 5% OFF all fees
Sign Up Now
BTC $71,807
Bull Run 2026 | 5% Off Fees Open your Binance account today
Sign Up

ChatGPT Share Link Flaw Let Hackers Read Private Chats

  • A malicious ChatGPT share link could execute attacker code in a victim’s logged‑in session, allowing private conversations to be read.
  • Researchers at Parameter reported the flaw to OpenAI on October 18; the company fixed it eight days later and rated it P1, its most severe category.
  • The attack exploited a MathLive listener in shared chats that accepted messages from any browser frame, bypassing the Preview sandbox.
  • OpenAI paid a $5,000 bounty for the disclosure; no real‑world exploitation was reported.

Researchers at security firm Parameter discovered that a manipulated ChatGPT share link could run arbitrary code inside any signed‑in user’s session, exposing private conversations. The flaw was reported through OpenAI’s bug bounty program on October 18 and fixed by October 26.

- Advertisement -

The attack did not require the victim to install anything or log in through the attacker. Clicking the link and hitting ChatGPT’s Preview button once was enough; a later version made Preview open automatically.

Once triggered, the attacker’s code accessed the victim’s recent conversations and could even open a private chat that had never been shared. “Essentially, we could perform any action on behalf of the victim’s ChatGPT account,” the researchers wrote.

The breach succeeded because a math library called MathLive, loaded into the main ChatGPT page, accepted keyboard‑layout messages from any sender without verifying the origin. Those messages could carry raw HTML, which the Preview frame used to inject an unsandboxed frame sharing the chatgpt.com identity.

ChatGPT’s content‑security policy allowed scripts from YouTube, enabling the final payload. OpenAI reproduced both the code execution and the private‑chat proof, then rolled a fix across two frontend versions by October 26.

- Advertisement -

Parameter traced the problem to three design choices: a listener that did not check message origin, a message that allowed raw HTML as keyboard data, and HTML that could create an unsandboxed frame. The fix blocks the messages the attack relied on, and users need to take no action.

✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.

Previous Articles:

- Advertisement -
Ad
Altseason Is Loading. Don't watch from the sidelines.
SOL $90.51
DOGE $0.0963
LINK $9.02
SUI $1.00
5% off fees when you sign up
Start Trading
Ad
Pay Less on Every Trade. For Life.
$10K/mo volume Save $60/yr
$50K/mo volume Save $300/yr
$100K/mo volume Save $600/yr
5% off all trading fees when you sign up
Claim Your Discount

Latest News

ImmunityBio founder touts sepsis trials, NantX after Binance theft

Patrick Soon-Shiong renewed his “NantX coming soon” teaser after a $6.3 billion Binance theft,...

Crypto.com Shuts Hong Kong App, Eyes Institutional Exchange

crypto.com will shut down its consumer app in Hong Kong on Feb. 1, 2027,...

Kim Yo Jong denies North Korea behind Binance hack, hints US

Kim Yo Jong denied North Korea's involvement in the $6.3 billion Binance hack and...

Strive buys 124 BTC, fourth among public companies

Strive acquired 124 Bitcoin for roughly $10 million between Oct. 24 and Oct. 30,...

Claude finds 9.8 zero-click Linux root bug

Claude discovered a zero-click Linux vulnerability (CVE-2026-98432) rated 9.8 out of 10, described as...

Must Read

This is How to Buy and Sell Bitcoin

Now more than ever, there are a variety of ways to enter and exit the crypto market. While this is good, the availability of...
Ad
Altseason Is Loading. These 4 coins are trending right now.
SOL $92.12
DOGE $0.0950
LINK $9.02
SUI $1.02
5% off spot fees when you sign up
Start Trading