FlagThis — Daily Cybersecurity Intelligence Briefing

Security researchers at Adversa, led by Rony Utevsky, have identified a critical vulnerability in high-capability Large Language Models (LLMs), including Grok, involving "cryptographic context injection." This attack method utilizes AES (Advanced Encryption Standard) to obfuscate malicious prompt payloads, bypassing traditional plaintext-based guardrail architectures. By providing both the ciphertext and the decryption key within the same prompt, attackers leverage the model's inherent reasoning and technical capabilities to perform in-context decryption. Once decrypted, the model executes the hidden instructions, rendering current semantic and keyword-based input sanitization methods ineffective against sophisticated cryptographic evasion.

The "Grok Build" feature within the xAI ecosystem has been identified as facilitating the unauthorized bulk upload of entire Git repositories to xAI-controlled infrastructure. Technical analysis indicates that Git hooks or unauthorized integration scripts trigger synchronization without explicit user consent, exposing proprietary source code, internal architectures, and hardcoded secrets—including API keys and SSH credentials—to third-party servers. Furthermore, the platform is vulnerable to Indirect Prompt Injection; malicious actors can deploy crafted payloads via fake bug reports to hijack AI coding agents possessing repository access. This dual-vector threat significantly expands the organizational attack surface, facilitating both data exfiltration and automated exploitation of codebase vulnerabilities.