On this page
Short answer: a well-run cloud tool can be more secure than something you run yourself. The question is not "cloud or not" but what the vendor does, what you configure, and what you cannot control. Use the 30-minute review below.
The real risks
- Account takeover. Weak or reused passwords without two-step verification are the most common way in.
- Over-shared links. Public links, guest access and forgotten shares expose more than people realise.
- Third-party integrations. Every connected app is another way in, and tokens are often over-permissioned.
- Vendor incidents. If the vendor is breached, so are you. How they respond matters.
- Jurisdiction. Where the data sits and who can compel access to it.
- Leaving. Getting your data out, and getting it deleted, when you end the contract.
What good looks like
| Area | What to expect |
|---|---|
| Assurance | An independent audit report (such as SOC 2 or ISO 27001), which you can request |
| Identity | Single sign-on, two-step verification, password policy and session control |
| Access | Roles scoped per project, read-only roles, guest access with expiry |
| Visibility | An audit log you can export and stream to your own tools |
| Data | Encryption in transit and at rest, a documented region, backups, a deletion process |
| Integrations | Scoped, revocable tokens with expiry and rate limits |
| Incidents | A stated process and a disclosure contact |
A 30-minute vendor review
- Ask for the audit report and read the scope. Check the date and what it covers.
- Check identity controls. Can you enforce SSO and two-step verification for everyone?
- Look at roles. Create a viewer and a member, and see what each can and cannot do.
- Test sharing. Create a guest link. Can it be limited and expired? Can you see and revoke all links?
- Find the audit log. Is it complete, filterable and exportable?
- Read the data terms. Where is the data stored, and who are the sub-processors?
- Check AI features. What is sent, to whom, and can you turn it off?
- Test the exit. Export your data. How complete is it? Ask how deletion works.
Write down the answers. If a vendor is vague on any of them, that is information.
When to choose self-hosted instead
Self-hosting moves the risks to you. It is a good fit when you have data residency rules, an air-gapped network, or a security team that wants full control, and you have people able to patch, back up and monitor the server. It is a poor fit if nobody will own that work. Our guide to self-hosted project management software covers the trade-offs.
NUSK's own controls
NUSK can be hosted by us or run on your servers. The controls are the same in both:
- Five roles enforced on every endpoint, per-project permissions and read-only guest links.
- Single sign-on with Google, Microsoft or any SAML 2.0 provider, SCIM provisioning, authenticator-app two-step verification, a password policy and login lockout.
- Signed-in device list with remote sign-out, and immediate effect when a user is deactivated.
- An audit log with filters, CSV export and streaming to a SIEM.
- Personal access tokens with read-only or read-and-write scope, expiry and rate limits.
- Webhooks signed with HMAC-SHA256.
- AI features that can be turned off entirely, with keys stored encrypted.
We do not claim a security certification we do not hold. If you need a certified audit report, ask us what exists today.
Talk to us about hosting · Start free
See it with your own project
Import a plan or a Jira project into a free NUSK workspace and judge it on your own data.