Capgo CVE-2026-88860

Exploitation requires that the attacker previously held a channel-specific permission override in Capgo and that this override was NOT removed when their last organization role binding was deleted - the stale override is the exact prerequisite. The attacker must retain a working Capgo account/credentials for the same principal (this is an offboarding/insider-adjacent condition, not anonymous access), and the target must have used per-channel overrides rather than relying solely on base org roles. What limits exploitation: the attacker must have been a prior privileged member with a channel override (a random external attacker cannot trigger it), and organizations that never granted channel overrides, or that fully purge accounts on offboarding, are not exposed.

The provider CVSS 4.0 vector (AV:N/AC:L/PR:N/UI:N/VC:H/VI:H/VA:H, 9.3) rates this critical, but the PR:N (no privileges) metric is questionable: exploitation inherently requires an account that previously held a channel override, so the attacker authenticates as a known former member - this is an authorization escalation for an existing/expired principal, not a truly anonymous remote attack, which argues for PR:L. That nuance lowers the practical exposure somewhat, since a random internet attacker cannot trigger it - it is an insider/offboarding risk. No EPSS score, KEV listing, or public POC is provided, so there is no evidence of active exploitation or automation; the signal here is a credible vendor/VulnCheck advisory rather than observed attacks. The impact is nonetheless serious where it applies: control over production OTA channels means an attacker could ship malicious or broken updates to end-user devices, so any organization that offboards developers/contractors from Capgo should treat stale-override cleanup as a priority, and a mid-sized mobile team using Capgo for production releases should verify that revoked users no longer retain channel access. Overall: genuinely high-impact for affected deployments, but exploitation is gated on prior insider access rather than being broadly wormable.

A contractor is granted a channel override to publish updates to a production OTA channel, then is offboarded and their organization role is deleted. Because Capgo leaves the channel override in place, the former contractor logs in with their still-valid account and pushes a malicious OTA version to production, silently delivering attacker-controlled code to all app users on that channel. No public exploit code is required - the flaw is triggered simply by exercising the retained permission through normal Capgo API/UI actions.