Hackers Exploit Critical Langflow and Ruby on Rails Flaws in Active RCE Attacks

Threat actors are actively exploiting two newly disclosed remote code execution vulnerabilities affecting Langflow and Ruby on Rails. These campaigns focus on cloud credential theft, host reconnaissance, and the establishment of command-and-control (C2) functions.

VulnCheck researchers have reported exploitation targeting CVE-2026-0768 in Langflow, a low-code platform for building AI-powered applications and automated workflows.

This vulnerability is an unauthenticated remote code execution flaw in Langflow’s custom component editor, which could allow attackers to execute commands on exposed servers without valid credentials.

Critical Langflow and Ruby on Rails Flaws

Caitlin Condon, VP of Security Research at VulnCheck, noted that the company’s Canary sensors recorded over 50 detection events shortly after exploitation began.

The malicious requests combined reconnaissance with credential-harvesting techniques, as attackers sought to obtain secrets commonly used in AI and cloud environments.

The observed payloads queried environment variables such as LANGFLOW_SUPERUSER, OPENAI_API*, AWS_ACCESS*, and AWS_SECRET*. If these variables are exposed, attackers could gain administrative access to Langflow, as well as access to OpenAI API keys, AWS access key IDs, and AWS secret access keys.

Attackers also attempted to read the file located at /root/.cache/langflow/secret_key, check SSH access, and assess the size of .bash_history.

This activity could help them identify stored application secrets, access paths, and command history that might reveal additional credentials or insights into internal infrastructure.

The exploitation traffic associated with Langflow primarily originated from Russia. It exclusively targeted VulnCheck Canary systems in the United Kingdom at the time of observation.

No public proof-of-concept exploit for CVE-2026-0768 is currently known, suggesting that threat actors may have independently developed exploit code or obtained it through private channels.

In a separate campaign, VulnCheck observed active exploitation of CVE-2026-66066, a Ruby on Rails Active Storage File Read-to-Remote-Code-Execution vulnerability. This activity was detected on Canary systems in Singapore, Israel, and the United Kingdom.

Unlike the broader Langflow activity, the Ruby on Rails exploitation was traced to a single source IP address in France. The attacker subsequently established command-and-control communications with a host in Israel, indicating that the campaign extended beyond initial exploitation into post-compromise operations.

These two campaigns highlight attackers’ increasing focus on internet-facing application frameworks and AI workflow platforms. Langflow deployments can hold high-value AI and cloud credentials, while Ruby on Rails environments may provide access to critical web applications and backend infrastructure.

Organizations should immediately identify exposed Langflow and Ruby on Rails systems, apply available vendor updates and mitigations, and review logs for any suspicious file reads, command execution, and outbound network activity.

Teams should also rotate potentially exposed AWS and OpenAI credentials, audit environment variables, restrict unnecessary public access, and monitor for unauthorized connections to unfamiliar command-and-control infrastructure.

Prevent incidents due to slow investigations. Power your Tier 1 with threat intelligence from 15K SOCs: Integrate TI Lookup in your SOC