Safeguard
Software Supply Chain•

How should an open source maintainer answer a vendor security questionnaire?

Maintainers owe no one a questionnaire. Publish a security policy, point to public evidence, keep one answers file, and ask the company to fund the extra work.

Safeguard Security Team
4 min read

An open source maintainer can answer a vendor security questionnaire with public evidence instead of bespoke paperwork. Publish a security policy, point to your repository settings and release process, keep one reusable answers file, and say plainly what you do not promise. If a company wants more, ask it to contribute time or funding.

Why are maintainers getting questionnaires?

Companies that depend on a package are often asked by their own customers and auditors to show they manage supplier risk. The easy way to do that is to send a form to the maintainer. The form was written for a commercial vendor with a sales team, a legal department and an SLA. A volunteer with a licence file has none of those.

Start from a clear position. Most open source licences supply the software as is, with no warranty. You are not a vendor to the company that downloaded your code, and a questionnaire does not change that. You can still choose to help, and a good set of public answers is the cheapest way to do it.

What public evidence can replace a form?

Build an answers page once. Most of the questions map to things you can show without writing new prose.

  • How do people report a vulnerability? GitHub documents adding a security policy: a SECURITY.md file that states supported versions and how to report. Link to it.
  • Is code reviewed? Point to your branch protection and review settings.
  • Are dependencies pinned and updated? Point to your lockfile and update automation.
  • Do CI tokens have limited permissions? Point to the workflow permissions in your pipeline files.
  • Are releases signed? State what you do, or say you do not.

The OpenSSF Scorecard tool runs automated checks on a public repository, including Pinned-Dependencies, Token-Permissions, Security-Policy, Signed-Releases, Branch-Protection and Code-Review, and prints a score with reasons. A published scorecard result gives a requester neutral, repeatable evidence. Treat the score as a conversation starter, not a grade you are obliged to defend.

How do you answer questions you cannot or will not meet?

Be literal. "Do you have a 24 hour incident response SLA?" The accurate answer is no, with what you do instead: security reports go to a private channel and get triaged when a maintainer is available. A false yes creates real exposure for you and for the asker. A calm no, plus the evidence of your process, is more useful to a risk reviewer than a form full of ticked boxes.

Questions about certifications, insurance, background checks, data centres and penetration tests usually do not apply to a code repository. Say "not applicable, this is a public source repository" and move on.

When should you ask for something back?

If a company needs custom attestations, a call, a signed document or a response deadline, that is consulting work. Offer the choices: use the public page, sponsor the project, or hire a maintainer. Many organisations already have a budget line for exactly this. A short, friendly template helps you apply the same rule to everyone.

How do you keep this from eating your week?

Keep a single file in a docs folder with your standard answers, dated and versioned. Update it when your process changes. Reply to every request with a link, and keep a short log of who asked, so you can see whether one company is sending the same form repeatedly.

If a request carries a link, an attachment or a login prompt from an address you do not know, check that it is real before you open it. Fake questionnaires are a way to get maintainers to run something or hand over credentials.

What to do

  1. Add a SECURITY.md that says how to report and what to expect.
  2. Run the Scorecard on your repository and fix the checks that are cheap to fix.
  3. Write an answers page with evidence links and honest gaps.
  4. Make a reply template: link, scope statement, and what extra support costs.
  5. Verify unknown senders before opening files or links.
  6. Review your own dependencies before you answer questions about them. Safeguard's free scan at gold.safeguard.sh/scan covers up to 100 dependencies.

Quick check

Do I have to complete a vendor questionnaire for a company that uses my library?

No. Open source licences generally provide the software as is. You can offer public evidence as a courtesy, and ask for sponsorship if they want more.

What is the most useful single file to publish?

A SECURITY.md that states how to report a vulnerability and which versions you support.

Sources

Last reviewed: 18 October 2026.

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.

Self-healing security runs on Safeguard.

Your first fix PR is minutes away.

No sales call required, even your agent can complete the purchase over MCP.