50% of CISOs Report Security Burnout. GitHub Copilot Reports First CVE
Big Picture
If you've been feeling burnt out recently you might take comfort in the fact that your boss (or your boss's boss) is stressed too. Nagomi Security's hilariously named "CISO Pressure Index" reported this week that 50% of CISOs feel burnt out and "it's affecting business readiness." You know what they say, with big titles comes big responsibilities. Jokes aside, the report highlights personal blame when breaches happen, staff reduction pressures, and dealing with sprawling tool stacks as reasons for the stress. Oh and of course AI.
Speaking of AI, GitHub Copilot just disclosed its first CVE via prompt injection, ransomware crews are shipping vibe-coded malware, the revised OWASP Top 10 added supply chain risk as a category. With so much going on it might explain why the Pentagon's flip to mandatory third party certification on CMMC 2.0 has received so little fanfare.
CISO Burnout Reaches Crisis Levels
Nagomi Security's 2025 CISO Pressure Index paints a grim picture: 80% of CISOs report high or extreme pressure, 87% say it's increased over the last year, and 40% have considered leaving their roles. The operational impact is measurable—44% say burnout is affecting their ability to prepare for breaches, and 73% experienced a major security incident in the last six months.
Tool sprawl and tool performance emerge as major stress contributors. 65% of CISOs oversee 20 or more security tools, yet 58% say breaches occurred despite having tools in place to stop exactly that type of attack.
Takeaways
AI contributes to this stress in two ways. First, 59% of CISOs cite agentic AI attacks as their top threat, but second, 82% report pressure to reduce staff using AI-driven automation tools. More AI attack surface, fewer people to defend, lack of confidence in tooling. Nice trifecta.
Patch Tuesday Cascade Creates Remediation Backlog
We're calling this week Patch Tuesday given the rush of CVEs, exploits in the wild, and patches that hit the newswire:
• Microsoft dropped 63 CVEs including a Windows Kernel zero-day actively exploited in the wild (CVE-2025-62215)
• Oracle pushed emergency patches
• SAP released critical fixes including a hardcoded credentials flaw so severe they're deprecating the entire product
• React Native CLI disclosed a critical RCE (CVE-2025-11953)
Total damage: 80+ critical and high severity CVEs across major platforms in seven days.
Takeaways
Patch Tuesday isn't unique anymore. The last four AppSec Weekly briefings have featured similar CVE volumes. This is the new normal—which explains both the CISO stress levels we covered above and why automated remediation topped everyone's radar after last week's OpenAI Aardvark announcement.
First GitHub Copilot CVE Exposes AI Coding Risks
AI, it is the best of times, it is the worst of times. GitHub Copilot Chat just earned its first CVE (CVE-2025-53773), joining the ranks of tools that have graduated from "helpful assistant" to "potential security liability." The vulnerability—a prompt injection attack that enables arbitrary code execution—works by embedding malicious instructions in repository files that Copilot reads. When developers interact with compromised code through Copilot Chat, the injected prompts trick the AI into modifying sensitive workspace configurations, ultimately executing attacker-controlled code.
The technical specifics are darkly funny: a case-sensitivity bypass on Windows file paths lets attackers modify .vscode/settings.JSON (uppercase) instead of the protected .vscode/settings.json (lowercase), sidestepping approval requirements. From there, it's a short hop to changing PHP executable paths and running arbitrary code.
This is the first CVE assigned to GitHub's flagship AI coding assistant—a tool used by roughly 14 million developers. That's a massive attack surface, and it validates concerns that have been simmering as AI tools continue to proliferate across development workflows.
Takeaways
The prompt injection vector affecting Copilot isn't unique to Microsoft's product. Any AI system that ingests external content is potentially vulnerable to adversarial inputs designed to manipulate model behavior.
If you're running GitHub Copilot (or any AI coding assistant), patch immediately and treat these tools like what they are: powerful automation with an expanding attack surface. And remember that "move fast and break things" is a lot less charming when the thing that breaks is your security posture.
CMMC Enforcement and OWASP Top 10 Update Drive Compliance Urgency
The Pentagon flipped the switch on CMMC 2.0 enforcement November 10th, moving thousands of defense contractors from self-attestation to mandatory third-party certification. If you're in the DoD supply chain, you now need to prove—not just document—your cybersecurity capabilities.
Meanwhile, OWASP dropped their 2025 Top 10 update with a new A10 category carved out specifically for supply chain risks, the first time dependency security and third-party code threats warrant their own distinct classification.
Takeaways
CMMC 2.0: Third-party assessors will look at your actual vulnerability remediation throughput, not your policy documentation. "We have a policy" won't cut it when you're sitting on a 100,000-item backlog with anywhere near the industry-average 252-day mean time to remediation (if this is you, we should talk about Pixee).
OWASP A10: Your scanner can find supply chain vulnerabilities all day. The bottleneck is triaging what's exploitable and fixing it without grinding your development pipeline to a halt.