Open Source Isn’t the Risk. Unexamined Software Is.

By Haines Eason

Nov 17, 2026

Share this post

hands typing on laptop keyboard

QUICK ANSWER: Can MSPs trust open-source software? Yes, when they evaluate it with a defensible process. Jared Casner, co-founder of Blacksmith InfoSec, advises managed service providers (MSPs) to review any open-source component that runs in a client environment, touches client data, affects availability, or supports service delivery. The same five-question framework applies to commercial software, too.

Open-source software comes without the familiar assurances MSPs expect from commercial vendors. There might be no SOC 2 report, security questionnaire, account manager or even a company standing behind the code.

But Jared Casner argues that does not make open source inherently less secure. It makes the work of evaluating it different—and exposes how much software risk MSPs might already be accepting without examining it.

Casner, co-founder of Blacksmith InfoSec and a software developer with more than 25 years of professional experience, expanded beyond open source itself at a ChannelCon presentation. His message: MSPs need a defensible process for understanding what software can do, what it touches and who will own the risk after it enters their environments.

“You’re all using open source software every day,” Casner said. “There’s open source everywhere in your life, whether you know it or not, and whether you’re paying attention to it or not.” That makes the question less about whether MSPs should trust open source and more about how they decide what to trust at all.

Accountability Is the Real Difference Between Open Source and Commercial Software

Casner pushes against easy assumptions on both sides of the open-source debate. Commercial software is not automatically safe simply because a vendor stands behind it. Contracts primarily protect vendors, he notes, and commercial products still experience vulnerabilities and security incidents. Open source, meanwhile, is not automatically safer simply because its code is readily available for inspection.

“More eyes can mean safer when it’s a very popular package,” Casner said. “But when it’s not that popular, you’re getting fewer eyes on it.” Nor does “free” necessarily mean inexpensive. An organization that self-hosts open-source software also assumes responsibility for infrastructure, uptime, maintenance and scalability.

The more important distinction is accountability. A commercial product gives an MSP someone to question. An open-source project might be maintained by a company, an active community or one exhausted developer working in their spare time. History shows how much could depend on that distinction. Casner points to Left-pad, the tiny JavaScript package whose removal from the npm ecosystem disrupted thousands of dependent projects, and XZ Utils, where an attacker spent years gaining trust before introducing malicious code. The vulnerability is not necessarily “open source.” It is dependency without sufficient visibility into the humans and software underneath it.

How AI Coding Tools Add Hidden Open-Source Dependencies

That problem is becoming more consequential as AI lowers the barrier to software development.

Casner asked how many attendees were using AI to build software—and how many had begun developing software after not doing so previously. Several hands went up. AI coding tools, he warned, can introduce open-source dependencies regardless of whether the person prompting the system understood what had been added.

“As you’re developing software, you are almost certainly injecting open source dependencies into your software, whether you know it or not, and whether you’re actually evaluating the risk that goes along with that or not,” Casner said.

The same principle applies well beyond AI-generated applications. A PowerShell script copied from Stack Overflow is open-source code, too. So is a GitHub gist or a package pulled into a larger project.

When Open-Source Software Needs a Security Review

Casner offered MSPs a simple threshold for deciding when those components deserved scrutiny: If open-source software runs in a client environment, touches client data, affects availability or becomes part of service delivery, it is in scope.

The level of review should then follow the potential consequence. A low-privilege internal utility does not warrant the same scrutiny as a configuration tool with administrative access to a Microsoft 365 tenant. An abandoned package should not receive the same confidence as a project with frequent releases, an active community and a documented security-disclosure process. The point is not to eliminate risk. It is to know which risk MSPs are accepting.

5 Questions MSPs Should Ask Before Approving Open-Source Software

Casner distills potential evaluations into five questions:

  • What does the open source tool do, what will it touch and what happens if it breaks?
  • Who maintains it, and is the project alive?
  • How will its maintainers handle security?
  • What dependencies and supply-chain risks come with it?
  • How will the MSP monitor it—and who will own it—after approval?

4 Possible Outcomes of an Open-Source Review

Those answers could lead to four outcomes: Approve the tool, approve it with guardrails, limit it to a pilot or sandbox, or reject and replace it. The framework also gives MSPs language for conversations that are otherwise difficult: Telling a client, technician or owner that a useful tool could be deployed, but only with limited permissions, network isolation or other safeguards.

The Same Framework Applies to Commercial Software

Then an audience member asked the question that widened Casner’s argument: How different was this process for proprietary software? “It really does not change,” Casner said.

Commercial software changes where MSPs look for answers. Instead of a security.md file in a GitHub repository, there might be a vendor security page, penetration testing, a bug-bounty program or a SOC 2 report. Instead of inspecting dependencies directly, an MSP might have to request a software bill of materials.

But the underlying questions remain. That makes Casner’s framework as much an indictment of unexamined commercial software as a method for trusting free tools.

Why Every Approved Tool Needs One Owner

The final question in Casner’s framework might also be the easiest to neglect. Approval is not the end of risk assessment. Software changes. Maintainers change. Dependencies change. New vulnerabilities emerge. Something that was defensible when adopted could become indefensible later. That requires monitoring—and, critically, at least one person responsible for it.

“An ownership problem exists when there are too few owners, meaning less than one owner,” Casner said. “And when you have too many owners, meaning more than one owner, you have an ownership problem.”

With no owner, nobody takes responsibility. With multiple owners, responsibility can become ambiguous. For MSPs, this can turn software evaluation from a purchasing decision into an ongoing operational obligation. Open source merely makes that obligation harder to ignore.

Frequently Asked Questions

Is open-source software safe for MSPs to use? Open-source software can be as safe as commercial software when it goes through a structured evaluation. The key difference is accountability. A commercial product has a vendor behind it. An open-source project might be maintained by a company, an active community, or a single volunteer developer.

When should an MSP review an open-source tool? An open-source tool is in scope if it runs in a client environment, touches client data, affects availability, or becomes part of service delivery. Match the depth of review to the potential impact. A tool with admin access to a Microsoft 365 tenant warrants more scrutiny than a low-privilege internal utility.

What are the possible outcomes of an open-source software review? There are four: approve the tool, approve it with guardrails, limit it to a pilot or sandbox, or reject and replace it. Guardrails can include limited permissions or network isolation. They give MSPs a clear way to say yes with conditions.

How do AI coding tools create open-source risk? AI coding tools can pull open-source dependencies into software whether or not the person prompting them knows it. The same applies to scripts copied from Stack Overflow or GitHub gists. Every one of these brings dependency risk that MSPs need visibility into.

Does the evaluation process change for commercial software? The core questions stay the same; only the sources of answers change. For commercial software, MSPs look at vendor security pages, penetration test results, bug-bounty programs, SOC 2 reports, and software bills of materials (SBOMs).

Who should own open-source software after it’s approved? Exactly one person. Software, maintainers, and dependencies keep changing, so approval starts an ongoing monitoring obligation. With no owner, nobody takes responsibility. With several owners, responsibility gets ambiguous.

Get more practitioner-to-practitioner insights and education at ChannelCon 2027, July 27-29, in Orlando. Register now.