Symantec Flags New Aoki Malware for Hijacking AI Coding Assistant Tokens

Symantec has issued an advisory about a newly identified variant of the Aoki infostealer, which specifically targets authentication credentials used by AI-driven coding assistants. The malicious code is capable of exfiltrating API keys and access tokens from services such as Anthropic's Claude and Google's Gemini, potentially giving attackers unauthorized access to developers' cloud resources and proprietary code generation tools.

The Aoki family of malware first appeared in 2022, primarily focusing on stealing banking information and login details for popular web services. This latest iteration reflects a shift in threat actors' objectives toward the burgeoning market for generative AI platforms, where API keys serve as the primary gatekeepers for paid usage and integration into development pipelines.

According to Symantec's analysis, the malware operates by scanning a victim's system for configuration files, environment variables, and credential stores that contain tokens for the targeted AI assistants. Once located, the data is packaged and transmitted to command‑and‑control servers controlled by the attackers. The stolen credentials can then be used to run large‑scale prompts, generate code, or even extract proprietary algorithms from the compromised accounts.

Industry observers note that the rise of AI coding assistants has accelerated the adoption of API‑based billing models, making the associated keys lucrative assets for cybercriminals. By hijacking these tokens, threat actors can accrue significant usage fees on victim accounts or leverage the services to automate further attacks, such as creating malicious code snippets that bypass traditional security checks.

Security researchers recommend several mitigation steps: developers should store API keys in secure vaults rather than plaintext files, rotate credentials regularly, and monitor usage patterns for anomalies. Additionally, endpoint protection solutions that can detect known Aoki signatures or suspicious file‑system activity are advised to be kept up to date.

Symantec's warning follows a broader trend of malware families adapting to the AI boom, underscoring the need for heightened vigilance as developers integrate these tools into their workflows. While no large‑scale breach linked to the new Aoki variant has been publicly confirmed, the advisory serves as a reminder that the security of AI service credentials is now a critical component of overall cyber hygiene.