Glossary / practitioner vulnerability-management governance

Vulnerability disclosure policy and security.txt

When a stranger finds a hole in your system, they need to know where to report it and that you will not sue them for doing so. A short public policy and a security.txt file give them both.

Researchers and members of the public regularly find flaws in websites and products they do not own. If there is no obvious place to report them, or the site only shows a stern warning against unauthorised use, many will not bother, and some will publish instead. A vulnerability disclosure policy (VDP) turns these chance finds into free early warning.

What the policy should say

CISA Binding Operational Directive 20-01 (2 September 2020) made a published VDP compulsory for US federal civilian agencies. It does not apply to Australian organisations, but its required content is a practical checklist for anyone:

Scope: which systems are covered. CISA required at least one internet-accessible production system or service to be in scope at the time of publication.

Rules of testing: which types of testing are allowed and which are not, plus a ban on passing any personal information found to a third party.

How to report: where reports go (for example a web form or email address), what information is needed to find and analyse the issue, and a clear statement that reports can be anonymous.

Legal safe harbour: a commitment not to recommend or pursue legal action against good faith research that follows the policy, and to treat that activity as authorised.

Expectations: when the reporter can expect an acknowledgement, a pledge to be as transparent as possible about remediation steps, and an issue date for the policy.

BOD 20-01 also said a policy must not require reporters to submit personal information. It also required agencies to be able to receive unsolicited reports, through a monitored security contact handled by staff who can triage them, before the policy was published.

security.txt (RFC 9116)

RFC 9116 defines security.txt, a small machine-readable text file that tells researchers how to contact you about vulnerabilities and where your policy is. CISA notes that using it is not required under BOD 20-01, but it can help people find who to report to.

Location: for web services the file goes at /.well-known/security.txt, must be served over HTTPS, and must have a Content-Type of text/plain with the utf-8 charset. A file only applies to the exact domain or IP address it was retrieved from, not to subdomains or parent domains, so each public host that matters needs its own file or redirect.

Required fields: Contact (one or more of a mailto:, tel: or https:// URI, listed in order of preference) and Expires (a date after which the data should be treated as stale; the RFC recommends less than a year ahead). Useful optional fields include Policy (a link to your VDP), Encryption (a key for sensitive reports), Acknowledgments and Canonical. The RFC recommends signing the file digitally and using Canonical to show where it is meant to live.

The RFC warns that stale information can be worse than no file: reports may go to dead or wrong contacts and expose issues to third parties. That is why Expires is mandatory.

CyberStack guidance (our own advice, not CISA's or the RFC's)

Publish the policy before you publish the file, and test the reporting route end to end. A security@ mailbox nobody reads is worse than none.

Put a calendar reminder well before the Expires date, and include security.txt in the change process for every new public domain, including marketing and campaign sites.

Write the triage path down: who validates a report, who owns the fix, what you tell the reporter, and when the issue moves from vulnerability handling to incident response.

Track a few numbers so the programme is real: time to first response, time to validate, time to fix, and how many valid reports are older than 90 days.

See also:

Fact source: RFC 9116 — A File Format to Aid in Security Vulnerability Disclosure (security.txt), CISA — BOD 20-01: Develop and Publish a Vulnerability Disclosure Policy (2 September 2020).