The productivity revolution promised by AI coding assistants has arrived. Developers are shipping features faster than ever, with tools like GitHub Copilot, Amazon CodeWhisperer, and Claude Code becoming as essential to modern development as Git itself. But beneath this velocity lies a troubling reality that every security leader needs to confront: we’re scaling security debt at unprecedented speed.

Version 2026.4.0 of the widely-used @bitwarden/cli npm package (78,000 weekly downloads) has been identified as malicious. The package contains a sophisticated multi-stage credential theft worm that explicitly names itself "Shai-Hulud: The Third Coming", a direct callback to previous Shai-Hulud supply chain campaigns, and targets developer credentials including SSH keys, cloud secrets, and even MCP configuration files.

Cyber-enabled crimes cost Americans nearly $21 billion in 2025, a 26% increase from the previous year, according to the FBI’s latest Internet Crime Report. Phishing, extortion, and investment scams were the most commonly reported attacks, with AI-related scams driving some of the costliest losses. Phishing was the top attack vector, with these attacks leading to more than $215 million in losses. Notably, AI-assisted business email compromise (BEC) attacks cost victims more than $30 million.

The surgeon is thirty-two minutes into a procedure. The ambient scribe pod listening to the operating room is mid-encounter — transcribing, retrieving prior chart context, drafting the operative note for post-op sign-off. At the same moment, your SOC gets an alert: anomalous tool invocation from that pod, elevated egress volume, behavioral deviation from the agent’s baseline.

As daily Jira users, the GitProtect Team experiences the platform’s pros and cons firsthand. A recent notification from Atlassian about data loss was a sharp reminder of digital fragility. Without GitProtect, our enterprise-grade backup solution, the loss could potentially turn into a critical data incident. Let’s see what happened and how we responded.

It’s Tuesday morning at a mid-size fintech. A customer-support workflow runs on CrewAI in production: a Triage agent reads tickets, a Records agent pulls customer history, a Remediation agent drafts and sends the reply. A user submits a ticket with a pasted error log containing an indirect prompt injection. Triage summarizes and delegates. Records, interpreting instructions embedded in the summary, pulls 2,400 customer records instead of one.