← Back to blog

Vulnerability Disclosure Policy: What to Publish and How

November 18, 2026
Vulnerability Disclosure Policy: What to Publish and How

A vulnerability disclosure policy (VDP) is a short public page that tells outsiders how to report a security flaw in your systems, what they may test, and what you will do with their report. For a startup, small business, school or charity, a one-page policy and a monitored email address are enough to start. The policy only works if you also find and fix the flaws yourself, which is where a CREST-accredited penetration test comes in. We cover both below.

Solusec
solusec.co.uk
Find Vulnerabilities Before They’re Reported
Solusec’s CREST-accredited experts provide penetration testing with clear, actionable findings to help strengthen your security posture.
Get a Pentest Quote

Key Takeaways

  • A VDP is a published set of rules for reporting flaws. It is not a bug bounty, and you can state plainly that you pay no rewards.
  • The NCSC's Vulnerability Disclosure Toolkit rests on three parts: a reporting route, a policy and a security.txt file.
  • A basic policy covers contact method, what to include in a report, what the finder can expect, and what is in and out of scope.
  • The Computer Misuse Act 1990 does not separate good-faith researchers from criminals, so your policy must set clear, narrow boundaries.
  • A policy invites reports. A penetration test finds the issues first, so the reports that arrive are fewer and less serious.

Table of Contents

What a vulnerability disclosure policy is

Someone outside your organisation sometimes spots a weakness: an exposed admin page, a login form that leaks data, an old plugin. If they cannot tell who to contact, they may say nothing, post it publicly or, worse, leave it for someone who will exploit it. A VDP gives them a safe, obvious route.

The NCSC, the UK's National Cyber Security Centre, explains that receiving reports mitigates two risks: flaws being found and exploited by attackers, and reputational damage from public disclosure when no reporting route exists. ETSI, the European standards body, makes a similar point in its 2022 report on coordinated disclosure: without a route, finders may go public, or problems may sit undetected until a serious incident.

Policy, programme and bug bounty

The three terms get mixed up, and searchers use "policy" and "programme" interchangeably. In practice:

  • A policy is the published page: the rules for researchers.
  • A programme is the policy plus the internal work behind it: who reads reports, how you check them, how you fix and reply.
  • A bug bounty pays rewards for valid in-scope findings, so it needs a reward budget and firm rules about what earns a payment.

According to the IANS glossary entry on VDPs, a VDP invites reports without promising payment, and stating up front that no rewards are offered is the simplest defence against "beg bounties", where people demand money for trivial findings. The NCSC's own example policy takes the same line: no monetary rewards.

What "coordinated" means

Coordinated vulnerability disclosure means the finder and the organisation work together: the finder reports privately, you fix, and public release is timed with both sides in agreement. The international standard behind this, ISO/IEC 29147:2018, sets out how organisations receive reports and publish remediation information. Its stated goals include reducing risk, limiting the harm and cost of disclosure, and setting expectations for cooperation. A related standard, ISO/IEC 30111, covers what happens inside the organisation between receipt and disclosure.

If you want a plain-English walk-through of setting one up, our page on vulnerability disclosure policy and management covers the service side.

Why a small organisation should publish one

You may think a VDP is for banks and government departments. The Bank of England and a university such as Chichester both publish one, but so can a ten-person software house.

The practical reasons are these:

  1. You hear about problems early. The NCSC's Ollie N, Head of Vulnerability Management, says quick, easy and secure reporting straight to the affected department speeds up triage and fixing, and cuts the risk of compromise.
  2. You look like you take security seriously. Customers, funders and procurement teams notice. ETSI lists this as a direct benefit.
  3. Law and codes of practice are moving that way. The NCSC notes that UK legislation will require smart-device manufacturers to provide a public point of contact as part of a disclosure policy. Its Software Security Code of Practice guidance asks software organisations to publish an effective disclosure process.
  4. You reduce legal confusion. A clear policy shows what you consider authorised, which matters given the law covered below.

If you build software for other companies, this also becomes a sales point. See our page on cyber security for software houses and agencies.

What to put in the policy

The NCSC says a basic policy should state how to make contact, secure communication options, what to include in a report, what the finder should expect, and what is in and out of scope. Here is each part in the order a reader meets it.

1. A promise

One or two sentences: you welcome good-faith reports and will not pursue people who follow the policy. HackerOne's guidance, built on the 2016 US NTIA template, lists five elements: promise, scope, safe harbour, process and preferences. Keep the promise honest. You cannot waive the law for anyone, so promise what you control: that you will not take action against people who act within your stated rules.

2. Scope

List the domains, applications and IP ranges that are covered, and say what is outside, such as third-party suppliers' systems or your cloud provider's infrastructure. If you run on AWS, the provider has its own testing rules; our guide to AWS penetration testing rules and scope explains them.

3. Permitted and forbidden testing

The NCSC example policy has a "must not" list worth borrowing. It bars breaking the law, accessing unnecessary or excessive data, modifying data, high-intensity or destructive scanning, denial-of-service attempts, disrupting services, social engineering, physical attacks and demanding payment for disclosure. It also asks researchers not to report non-exploitable, best-practice-only matters such as missing security headers.

Its "must" list asks researchers to follow data protection rules, not share or redistribute any data they retrieve, and securely delete it once it is no longer needed or within one month of the fix, whichever comes first.

4. How to report

Give a dedicated address such as security@yourdomain, or a secure web form. The NCSC toolkit recommends a security.txt file published under /.well-known/ on all domains and subdomains. This is a small text file that machines and people can read, listing your contact route and policy link, so a researcher does not have to hunt for it.

Administrator configuring a website security contact file

Reports can contain sensitive detail about your weaknesses, so restrict who can read that mailbox and keep records tidy. The same discipline applies to any sensitive data; this piece on operational controls for protecting biometric data shows how access and retention controls are framed in another field.

5. What to include in a report

Ask for the affected website, IP or page, a short description of the type of flaw (for example "XSS vulnerability") and harmless, non-destructive steps to reproduce it. The NCSC example policy asks for exactly these, which speeds triage and reduces duplicate or malicious exploitation.

6. Response expectations

Set times you can actually meet. The NCSC example commits to responding within 5 working days and triaging within 10, and asks reporters to limit status enquiries to once every 14 days. A government department has a team behind that. A three-person charity should choose dates it can keep, because a missed promise annoys finders more than a modest one.

7. Disclosure and thanks

State that you do not pay rewards, that you will tell the finder when the fix is live, invite them to confirm it, and agree any public release together. Offering public thanks is optional and costs nothing.

A short template outline

SectionOne-line content
PromiseWe welcome good-faith reports and will not take action against those who follow this policy
ScopeNamed domains and applications; third-party systems excluded
RulesNo data modification, no denial of service, no social engineering, stop at proof of access
Reportingsecurity@ address or form, link to security.txt
Report contentsLocation, type of flaw, safe reproduction steps
Our responseAcknowledge within X working days, triage within Y
RewardsNone offered
DisclosureAgreed jointly after the fix

The NCSC toolkit includes a full example policy to base your own on, so you are adapting a tested document, not starting from nothing.

Handling a report once it arrives

A published policy with nobody reading the mailbox is worse than none. The NCSC toolkit sets out the response steps:

  1. Reply promptly and thank the finder.
  2. Pass the report to whoever owns the affected product, or discuss with the third-party supplier's manager.
  3. Do not force the finder to sign a non-disclosure agreement.
  4. Ask politely for more information if you need it.
  5. Confirm the issue is being managed.
  6. Send periodic updates.
  7. Confirm the fix and invite the finder to retest.
  8. Consider publicly acknowledging them.

Assign one named person and one backup, including during holidays and school breaks. Decide beforehand how you would rate a report as urgent. Our guide to vulnerability prioritisation for UK teams helps with that, and the pattern of what turns up in practice is covered in what we find on West Midlands networks.

If a report turns out to be a live compromise rather than a weakness, you need an incident plan rather than a patch. That is the job of incident response.

This is the part most templates skip, and it matters for a UK policy.

The Computer Misuse Act 1990 makes unauthorised access to a computer an offence under section 1. As Computer Weekly reports, the Act does not distinguish between malicious hackers and good-faith researchers, and in September 2026 the government rejected an amendment that sought a review of a statutory defence. Lord Clement-Jones told the Lords Grand Committee that leading UK cyber security companies are routinely forced to prohibit proactive vulnerability research on UK-based infrastructure because the legal risks are unacceptable.

A government plan for a researcher defence was reported by The Record in May 2026 as limited. It would cover only prosecutions over scanning internet-facing systems, would require researchers to stop the moment a vulnerability is identified, and would be restricted to British nationals with UK Cyber Security Council accreditations. Experts quoted said it would protect almost no one.

What this means for you:

  • Your policy is the clearest statement that testing within your rules is authorised by you. Write the scope precisely.
  • Be realistic about the stop point. The Record notes that asking researchers to stop the instant a flaw is identified makes the disclosure nearly worthless, because owners usually need proof the flaw is genuine. Define a sensible stopping point, such as proof that access is possible without downloading or altering data.
  • Practical Law's UK checklist on bug bounty and disclosure programmes names the Computer Misuse Act as the key legal risk for both sides, and covers a clear policy, reducing the risk of finders demanding money, and a formal handling process.
  • Have a solicitor review the final wording, particularly the safe harbour language and any data protection points.

A policy invites reports, testing finds the flaws first

Here is the point many small organisations miss. A VDP is a front door for other people's findings. It does not find anything itself, and unpredictable volunteers are not a testing schedule. If the first person to look hard at your website is a stranger, you have no control over when, how deeply or with what intent.

A penetration test reverses that. A scoped, authorised test finds the weaknesses on a date you choose, with a written report and fix guidance, before you publish an invitation to the internet. Our pages explain the difference between a penetration test and a vulnerability scan, and vulnerability management covers the ongoing side. The research we publish on vulnerability research and CVEs shows the kind of flaws that exist in real products, such as a local file inclusion in a monitoring panel.

A sensible order for a small organisation:

  1. Test first, fix the findings.
  2. Publish the policy and security.txt.
  3. Retest on a regular cycle, and after major changes.

Where Solusec fits

We are a CREST-accredited testing company, and you deal directly with the testers and assessors, not a sales team. Findings are written in plain English with recommendations sized to your organisation. Published prices: an external IP test is £3,000 one-off and a CMS website test is £2,500 one-off.

If a customer, funder or insurer is asking about your security at the same time, Cyber Essentials certification costs from £320 to £600 one-off, a gap analysis is £300, and an urgent turnaround is an extra £100. Schools can see Cyber Essentials for schools, and organisations that need the independently tested level can read about Cyber Essentials Plus.

Solusec

If you need a test quickly, tell us your deadline when you get in touch and we will scope around it. For the full range, start at solusec.co.uk.

FAQ

What is a vulnerability disclosure policy?

It is a published set of rules telling security researchers how to report flaws in your systems, what they may test and what you will do in response. It is not a bug bounty, so it need not offer payment. The NCSC's toolkit pairs it with a reporting route and a security.txt file.

Is a vulnerability disclosure programme the same as a policy?

Not quite. The policy is the public page. The programme is the policy plus the internal process behind it: who triages reports, how you fix issues and how you reply. A well-considered policy is the basis for a complete programme.

How quickly should we respond to a report?

Choose times you can meet. The NCSC's example policy commits to a response within 5 working days and triage within 10 working days. Smaller teams can set their own figures, provided a named person watches the mailbox.

Do we have to pay people who report a vulnerability?

No. A disclosure policy can state that no rewards are offered, as the NCSC's example does. Only a bug bounty programme promises payment, and it needs a budget and clear rules for what qualifies.

Sources

This article was created with assistance of AI. We never use AI with your data unless given explicit permission.