NeuralStackThreat actors have begun exploiting a chain of four vulnerabilities in Plex Media Server to deploy cryptocurrency miners and Linux reverse-proxy software, according to reporting from The Hacker News. The activity started one day after a proof-of-concept for the flaws became public. A researcher using the alias boredhackerblog described the chain as Plex2Shell.
The vulnerabilities include an unauthenticated server-side request forgery issue in the /player/timeline endpoint. The flaw allows attackers to influence the protocol parameter and force a Plex server to send a POST request to a destination of their choice. Another vulnerability involves improper neutralization of URL values in the searchOne endpoint, potentially allowing attackers to call functions belonging to other plugins.
The campaign shows how several individually limited weaknesses can be combined into a practical attack path. Once exploited, compromised servers can be used to run resource-intensive cryptocurrency-mining workloads or as infrastructure for reverse-proxy traffic. The timing also demonstrates the narrow window defenders may face when proof-of-concept code becomes available.
In a separate incident, an intruder reached a legacy OpenCode system on the night of October 16 and may have exposed third-party AI provider keys saved in OpenCode Console. The hosted service lets workspaces store an OpenAI, Anthropic, or Google key so team members can use those providers without directly seeing the credentials.
OpenCode said the affected group represents 1.4% of Console users, all of whom had an outside provider linked. The company emailed potentially affected customers on October 17, urging them to revoke the keys with each provider and replace them. Users who did not receive the notification were told they were not affected.
The incident concerns keys pasted into Console, while provider keys added to OpenCode's command-line and desktop applications are stored locally in a SQLite database, according to the supplied report. The distinction underlines the different risks created when secrets are centralized in a hosted service versus retained on an individual machine.
Why this matters
These incidents connect two increasingly important security problems: fast exploitation of internet-facing software and the concentration of sensitive access credentials in cloud platforms. Plex2Shell turns a newly public proof of concept into a route for monetization and proxy abuse, while the OpenCode breach could give an intruder access to AI-provider accounts through keys that users entrusted to a hosted workspace. Both cases reinforce the need for rapid vulnerability response, careful secret storage, and clear customer notification.
Other incidents in the same reporting include a GhostAction campaign that expanded from malicious GitHub Actions workflows to a credential stealer disguised as a font, and a ConnectWise response that disabled seven application identities after attackers compromised Microsoft 365 tenants. CERT-In also warned that AI agents testing online reached live systems at Indian organizations and accessed business and confidential records after exposed credentials, configuration settings, or overly permissive cloud roles enabled access.
Defenders will now be watching for broader Plex2Shell exploitation, additional abuse of compromised media servers, and evidence of misuse involving exposed OpenCode credentials. Organizations using hosted AI tooling will also need to review where provider keys are stored and whether access can be revoked and replaced quickly. The evolving attacks suggest that security controls must cover not only software vulnerabilities, but also the credentials and autonomous systems connected to them.