Threat Intel
PhantomRaven: LLM‑Generated JavaScript Stealer Used by Bug‑Bounty Threat Actor
CrowdStrike uncovered PhantomRaven, a JavaScript‑based information stealer crafted by a financially motivated bug‑bounty hunter. The tool harvests credentials and session data from web applications and is distributed via malicious scripts. Defenders should block known delivery URLs, enforce strict CSP, and monitor for anomalous outbound traffic.
What Happened
CrowdStrike identified PhantomRaven, a JavaScript‑based information stealer generated by a large‑language‑model (LLM) and distributed by a financially motivated threat actor who works as a bug‑bounty hunter. The script is embedded in malicious web pages and captures user credentials, session cookies, and other sensitive data before exfiltrating it to the attacker’s command‑and‑control server.
Impact
Any organization that hosts or interacts with web applications vulnerable to script injection or that relies on third‑party JavaScript libraries can be compromised. The stolen data can be used for credential stuffing, lateral movement, or sold on underground markets.
Mitigation Steps
- Block or quarantine known PhantomRaven delivery URLs and domains in web‑filtering or next‑generation firewall appliances.
- Enforce a strict Content‑Security‑Policy (CSP) that disallows inline scripts and only permits scripts from trusted origins.
- Deploy endpoint detection and response (EDR) solutions that flag anomalous outbound traffic to unfamiliar IPs or domains.
- Regularly audit third‑party JavaScript libraries and remove unused or outdated code.
Detection Tips
Look for unexpected outbound connections from browsers or web servers to unfamiliar IP ranges, especially those using HTTPS with self‑signed certificates. Monitor for high volumes of credential‑related API calls or sudden increases in failed login attempts that may indicate credential harvesting.