Google Chrome CVE-2026-106301

Exploitation requires the attacker to have ALREADY compromised Chrome's sandboxed renderer process (e.g., via a separate renderer RCE) - this is the primary gating precondition and makes the bug a chain link, not a standalone remote exploit. Given that foothold, the attacker must get the renderer to load a crafted HTML page (user interaction / page navigation, per UI:R) that abuses the Contextual Tasks feature to act as a confused deputy and cross into a privileged page. Limiting factors: affected only on Chrome versions prior to 155.0.8059.39, requires the prior renderer compromise, and requires the crafted page to be loaded, all of which sharply narrow realistic standalone exploitation despite the high CVSS.

Signals conflict in an important way. The input CVSS of 8.1 (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N) looks network-exploitable and unauthenticated, but the description explicitly states the attacker must have 'already compromised the renderer process' - so this is a sandbox/privilege-boundary escape that is a link in an exploit chain, not a standalone remote threat. The PR:N and AC:L in the supplied vector therefore understate the real-world precondition (a prior renderer RCE); verify scoring intent with Google. Supporting signals point to low operational urgency: EPSS is 0.18% (7th percentile), there is no CISA KEV listing, SSVC reports Exploitation: none and Automatable: no, and no public POC is referenced - this is 'no public exploit identified at time of analysis.' The one amplifying signal is SSVC Technical Impact: total, consistent with C:H/I:H. Net: a genuinely serious capability (full confidentiality/integrity bypass into a privileged page) but gated behind a prior renderer compromise and currently showing no exploitation evidence, so it is a patch-on-normal-cycle item rather than an emergency, tempered by Chrome's ubiquity.

An attacker who has already achieved code execution inside Chrome's sandboxed renderer (via a separate bug) lures or directs the compromised renderer to load a crafted HTML page that abuses the Contextual Tasks confused-deputy flaw. This tricks a more-privileged browser component into acting on the attacker's behalf, bypassing system access restrictions and reaching a privileged page. No public exploit code is identified at time of analysis; the user-interaction and prior-renderer-compromise requirements make this a chain component rather than a one-shot remote exploit.