Reported by 2 outlets
Critical Microsoft Copilot CoSnitch Flaw Lets Hackers Steal Sensitive Data With One Click
Why this score
- Breadth
- 1.00 / 2.0 · 2 sources
- Authority
- 1.20 / 2.0 · best outlet weight
- Severity
- 1.40 / 5.0 · CVSS + content signals
- Velocity
- 0.00 / 2.0 · updates in 48h
- Pickup speed
- 0.58 / 0.6 · time to second source
- Recency
- × 0.00 temporal multiplier
- Community
- + 0.00 / 1.5 · reddit signal
- Score
- 0.0 / 10 = base × recency + community
Timeline
-
original reporting
Critical Microsoft Copilot CoSnitch Flaw Lets Hackers Steal Sensitive Data With One Click
gbhackers.com · 56d ago · August 19th, 2026
A critical one-click vulnerability in Microsoft Copilot Personal, tracked as CVE-2026-24301 and dubbed CoSnitch. This flaw could enable an attacker to trigger malicious Copilot prompts, access data from connected OAuth applications, and silently transmit that information to an attacker-controlled server. Microsoft addressed this issue on August 18, 2026, following Varonis’s responsible disclosure in December 2025. […] The post Critical Microsoft Copilot CoSnitch Flaw Lets Hackers Steal Sensitive Data With One Click appeared first on GBHackers Security | #1 Globally Trusted Cyber Security News Platform .
-
update from cyberpress.org
CoSnitch Microsoft Copilot Flaw Enables Silent Data Theft With a Single Click
cyberpress.org · 55d ago · August 19th, 2026
Lior Adar found no evidence that CoSnitch was exploited in the wild before the fix; the attack chain highlights the data-exposure risks created when AI assistants have broad access to email, cloud storage, calendars, and other third-party services. CoSnitch Microsoft Copilot Flaw CoSnitch resulted from three separate weaknesses that could be chained together to execute attacker-controlled prompts, access linked accounts, and exfiltrate data without meaningful user interaction. The first issue involved an undocumented URL parameter used alongside Copilot’s standard ?q= query parameter. Attackers could craft a link that caused a prompt to run automatically as soon as the victim’s browser loaded the Copilot page. The victim did not need to type a prompt, approve an action, or interact with the page beyond opening the malicious URL. Once triggered, the prompt could instruct Copilot to query data available through connected applications, including Gmail, Google Drive, and Google Calendar. It could then use Copilot’s built-in URL-fetching functionality to transmit retrieved information to an attacker-controlled server. The exfiltration techni