Vulnerability disclosure policy

How to report a security vulnerability in AccessPoint or the Realizer platform, and what we will do about it.

Report a vulnerability to security@realizer.io. This mailbox is for security reports only — it is not customer support.

Realizer Services Inc. publishes this policy for AccessPoint and the Realizer platform. Last reviewed 2026-11-07.

1. Why this policy exists

AccessPoint handles access-to-information and privacy case files for government organizations. If there is a weakness in it, we would rather hear it from you than read about it later. This policy tells you what we run, what you may test, how to report what you find, what we will do, and how quickly.

We follow the coordinated vulnerability disclosure practices set out by CISA, the NSA, JPCERT/CC, NCSC-NL and NCSC-UK. We do not require non-disclosure agreements, and we do not ship silent fixes: when we fix a reported vulnerability, we say so in an advisory.

2. Scope

2.1 In scope

WhatWhere
The Realizer platformhttps://api.realizer.io and its stage service https://stage-api.realizer.io — licence validation, API-address discovery, the Teams tab router, the Teams notification relay, jurisdiction pack download, the marketplace landing page and webhook, the Finish Setup page, the SharePoint package install page, the hosted AccessPoint control plane and its billing webhook, these security pages and the advisory feed
Hosted AccessPointhttps://app.realizer.io (the entry, which forwards to a region) and the Canadian region https://ca.app.realizer.io with https://api.ca.app.realizer.io; the stage service https://stage-app.realizer.io, https://ca.stage-app.realizer.io and https://api.ca.stage-app.realizer.io — only inside organizations you create yourself (section 3.4)
The Training Labhttps://learn.realizer.io — as a learner, only in a class you were given the code for; ask us for a test class (section 3.4)
The AccessPoint demohttps://demo.realizer.io (it runs in your browser on made-up data and calls no AccessPoint service)
The Realizer admin consolehttps://admin.realizer.io — sign-in is for Realizer staff only, so findings that need no account
The Realizer websitehttps://realizer.io
The artifact hosthttps://get.realizer.io
The container registryrealizer.azurecr.io (the AccessPoint document-worker, media-worker and accessibility-worker images and the connector images)
The AccessPoint software we distributeThe API and portal packages, the SharePoint Framework package, the Teams app, the ARM/Bicep template, the deployment script, the connector packages and SDK, jurisdiction pack content

Testing the product itself. AccessPoint comes in two editions. The Microsoft 365 edition deploys into the operator's own Azure subscription and Microsoft 365 tenant: to test it, deploy your own instance and test that. We will help you get one running — ask at security@realizer.io. Findings from your own instance are fully in scope, including the deployment template, the database schema, the API, the portal, the web part and the worker image. Hosted AccessPoint is the same software run by us for many organizations: you may test it with a free trial, only inside organizations you create yourself and with accounts that are yours (section 3.4).

We do not limit participation by nationality, age or affiliation, and we accept anonymous reports. We may be unable to work with anyone we are legally barred from dealing with.

2.2 Out of scope

  • Any AccessPoint deployment operated by one of our customers. Those run in the customer's own Azure subscription and Microsoft 365 tenant, hold real case files about real people, and we cannot authorize you to test them. Do not probe a customer's instance. If you believe you have found a weakness in a specific organization's deployment, report it to that organization; you may also tell us (section 2.3).
  • Other organizations in hosted AccessPoint, and other training providers' classes in the Training Lab. They hold real records and courses that belong to someone else, and we cannot authorize you to access them. A weakness that reaches from one organization or class into another is exactly what we want to hear about: show it between two organizations or classes of your own, stop at proof (section 3.2) and tell us.
  • Microsoft services (Azure, Microsoft 365, Entra ID, Microsoft Graph, the commercial marketplace). Report those to the Microsoft Security Response Center at https://msrc.microsoft.com/report.
  • Third-party components we do not control, including the open-source libraries AccessPoint builds on (for example PDF.js, LibreOffice and Chromium). Tell us anyway if it affects AccessPoint users — we will coordinate with the project.
  • Findings without a security impact, including: missing security headers with no demonstrated exploit; weak TLS ciphers with no demonstrated attack; SPF, DKIM or DMARC configuration reports on domains that send no mail; rate-limit observations on anonymous endpoints without a demonstrated impact; version-banner disclosure; issues that require an out-of-date or unpatched browser; clickjacking on pages with no state-changing action; self-XSS; social-media account takeover claims; automated scanner output with no analysis.
  • Denial of service, load testing, and physical or social-engineering attacks against Realizer, our customers, our staff or our suppliers (section 3.3).

2.3 If your finding is about a customer's deployment

Tell the organization that operates it. If you tell us instead, we will pass it on through the one address we hold for that organization — its account contact — and through a security advisory inside their own AccessPoint; we hold no list of their security staff. We will not identify you to them without your permission, and we will not test their environment ourselves without their written authorization.

Hosted AccessPoint and the Training Lab are ours to run, so tell us directly. If a finding shows that an organization's or a training provider's data was reached, we handle it as an incident and tell them ourselves; we will not identify you to them without your permission.

3. Rules of engagement

3.1 Test broadly, exploit minimally

We would rather you used the techniques a real attacker would use than worked around an arbitrary boundary. In exchange, stop at proof: do only as much as is necessary to demonstrate that the vulnerability exists.

3.2 Protect data

If you encounter personal data, credentials, case content or anything else that is not yours, stop, do not save it, and tell us immediately. Do not access more records than you need to show the problem. Delete anything you retrieved once your report is acknowledged, and tell us that you have.

3.3 Do not

  • Introduce malware, or leave any code or account behind.
  • Copy, modify, encrypt or delete data in a system that is not yours.
  • Make configuration changes to our systems.
  • Access accounts, mailboxes or tenants that are not yours.
  • Repeatedly access a system beyond what your testing needs, or share your access with anyone else.
  • Run brute-force or credential-stuffing attacks.
  • Run denial-of-service or load tests.
  • Use social engineering, phishing, or physical intrusion against Realizer, our customers or our suppliers.
  • Publish before we have agreed a disclosure date (section 6).

3.4 Hosted AccessPoint and the Training Lab

These services are shared by many organizations, so a few more rules apply:

  • Test only organizations you created, with Microsoft accounts that are yours. Do not invite yourself into, sign in to, or read anything of an organization you did not create. Show a cross-organization weakness between two organizations of your own.
  • No load or volume testing: the service you would slow down is everyone's (section 3.3).
  • E-mail sending limits are part of the service. Do not use letters, invitations, confirmations or notices to send mail to addresses that are not yours, and reaching a sending limit is not a finding.
  • Do not make real payments to test billing; billing runs through our payment provider, which is out of scope.
  • In the Training Lab, the proof of work and the limits on joining are deliberate. Guessing class codes or resume codes is a brute-force attack (section 3.3); a weakness in the codes themselves can be shown in a class you were given.
  • When you are done, tell us which organizations you created, so we can remove them.

4. How to report

Email security@realizer.io. This mailbox is separate from customer support and is monitored for security reports only.

Encryption. If you would rather not send details by plain email, ask us for another channel and we will arrange one.

Anonymous reports are accepted. You will not be asked to identify yourself, though we cannot credit you (section 7) or tell you about the fix if we cannot reach you.

A report we can act on contains, at minimum:

  • what the issue is, in one or two sentences;
  • where it is (URL, endpoint, package and version, or file and line);
  • how to reproduce it, step by step;
  • what an attacker gets out of it.

An ideal report adds: a minimized proof of concept, your assessment of the impact and who is affected, a CVSS vector if you use one, any logs or timestamps of your testing (so we can tell your traffic from an attacker's), and how you would like to be credited.

Please write in English or French.

5. What you can expect from us

StageOur target
Acknowledgement that a human has your report2 business days (3 at the outside)
Triage decision — is it valid, what severity, are we fixing it10 business days of acknowledgement
Progress updatesAt least every 14 days while the report is open, without you having to ask
Fix or mitigationCritical 7 days, High 30 days, Medium 90 days, Low with the next scheduled release, from the triage decision
Advisory publishedWith the release that carries the fix, or sooner if there is a mitigation you should tell people about
DisclosureCoordinated with you (section 6)

We will tell you if a target is going to slip, and why, before it slips. If we decide not to fix something, we will tell you that too, with our reasoning, and you are free to say so publicly under section 6.

Severity is assessed on the impact to the operators of AccessPoint and the people whose information it holds, not on how hard the bug was to find.

6. Coordinated disclosure

  • Default: 90 days from our acknowledgement, or the day the fix is published, whichever comes first.
  • We will agree an earlier or later date with you where it makes sense — for example, more time for a fix that needs a schema migration in every customer's database, or less time for something already being exploited.
  • If a vulnerability is being exploited, we will move immediately and will not hold a fix or an advisory for an embargo.
  • Long embargoes do not make anyone safer; the more people who know, the more likely a leak.
  • We will not ask you to sign a non-disclosure agreement as a condition of reporting, and we will not use one to keep a fixed vulnerability quiet.
  • Our advisories are published on a public page with no login and no paywall.
  • If we and you cannot agree, you may publish at the end of the agreed window. We would rather coordinate than argue.

7. Credit

We name reporters in the advisory when they want to be named, in whatever form they ask for (name, handle, organization, or nothing). Tell us in your report.

We do not run a paid bug bounty today. If that changes, this policy will say so before it takes effect.

8. Safe harbour

If you make a good-faith effort to comply with this policy during your security research, Realizer Services Inc. will consider your research authorized, will work with you to understand and resolve the issue quickly, and will not initiate or recommend legal action against you in relation to that research. If a third party brings an action against you for activity carried out in accordance with this policy, we will make this authorization known.

This authorization is ours alone to give. It does not cover, and we cannot give authorization for:

  • organizations and training providers other than you in services we run: in hosted AccessPoint and the Training Lab your research is authorized only within organizations you created and classes you were given (section 3.4);
  • systems operated by our customers, including any AccessPoint deployment in a customer's own Azure subscription or Microsoft 365 tenant (section 2.2);
  • Microsoft's services and infrastructure, which are governed by Microsoft's own terms;
  • any activity outside this policy, including the prohibited activities in section 3.3.

Nothing here limits the rights of any third party, and this authorization does not override any law. If you are unsure whether something is in scope, ask us first at security@realizer.io — we would rather answer a question than read about a misunderstanding.

9. security.txt

Our machine-readable security contact is published at /.well-known/security.txt, as RFC 9116 requires, on both realizer.io and api.realizer.io.

10. Advisories

Published advisories are listed at /security/advisories, with an Atom feed at /security/advisories/feed.atom. There is no login and no paywall. Customers running an affected version of AccessPoint 2.5.0 or later also see the advisory inside AccessPoint, and those who have registered security contacts there receive it by email from their own organization's mailbox.

Service events — outages, degraded service and maintenance of these services — are posted on our status page at https://www.realizer.io/status, which runs on our website, separately from this platform, so it stays reachable when the platform is not.