The vulnerability in Uber Kraken agents arises from improper validation of integrity check values. Kraken agents download blobs peer-to-peer and validate pieces using CRC32 checksums rather than verifying the full SHA-256 digest of the entire blob before committing it to the content-addressable cache. Attackers positioned on the agent-to-agent path or acting as malicious peers can exploit this by supplying substituted content with forged CRC32 corrections that pass piece-wise validation. This results in cache poisoning with attacker-controlled container image layers or manifests, which are then propagated and executed by other hosts relying on the poisoned cache.

An attacker able to intercept or participate in the peer-to-peer blob distribution can inject malicious container image layers or manifests into the Kraken cache. This cache poisoning can lead to execution of attacker-controlled code on other hosts that pull from the poisoned cache, potentially resulting in remote code execution or system compromise. The vulnerability has a high attack vector (network), low attack complexity, no required privileges or user interaction, and impacts confidentiality, integrity, and system availability with high scope and impact.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is released, users should consider restricting or monitoring peer-to-peer blob distribution channels to trusted entities only and avoid using vulnerable Kraken agent versions in untrusted network environments.