Checking the software your business depends on
Map software suppliers and dependencies, verify update routes and access, and plan tested recovery when a component fails or needs an urgent security fix.

Your business software may depend on libraries, hosted services, integrations and update systems supplied by other organizations. A failure or compromise in one dependency can affect the work you do with the application. In July 2024, a faulty update from the security vendor CrowdStrike crashed about 8.5 million Windows devices, by Microsoft’s estimate, and disrupted airlines, banks and many other businesses.
Reliability and security need separate checks. A faulty update isn’t necessarily a cyberattack, and a signed package may still be wrong for your setup.
Map the dependencies that matter to work
Identify essential applications, their suppliers, integrations and owners. Record where updates come from and which accounts or services can install them.
For software you build, ask developers to keep a component inventory, including versions and sources. A software bill of materials can help with that, but it won’t show whether those components contain vulnerabilities or malicious code.
Include external services that don’t install anything locally. Authentication, APIs and hosted storage can still be critical dependencies.
Ask suppliers about development and response
Request their vulnerability-reporting route, support period, update process and incident contacts. Ask for evidence about the product you use, such as its release notes, security advisories and published support policy.
NIST’s Secure Software Development Framework offers a vocabulary for discussing secure development practices. Use it to frame your questions. A supplier that mentions a framework may not follow it, so ask what they’ve put in place.
The supplier-selection guide covers the broader assessment and onboarding.
Protect the route that produces and installs software
Limit access to source repositories, build systems, package publishing and deployment accounts. Protect credentials and record who approves changes.
Install packages from known sources and use integrity checks where they’re available. Investigate an unexpected update route or request to disable protection. Review supplier access when the work ends or an account changes ownership.
Separate routine users from administrative or deployment privileges where practical. A compromised ordinary account shouldn’t automatically control software distribution.
Match update testing to the risk
Use a representative trial group where the product permits staged deployment. Test business functions, integrations and recovery before wider rollout. Record which versions were accepted.
Balance testing with the exposure from delaying a security fix. The Cyber Centre’s patching guidance supports a managed assessment and change process. A fixed waiting period won’t suit every update.
Some hosted services and update channels don’t provide customer-controlled staging. Document that limit and plan continuity accordingly.
Prepare recovery with the supplier
Ask whether rollback is supported, what configuration is needed and how to retrieve data if the service is unavailable. Test recovery without assuming an older version remains safe or supported.
Keep verified supplier contacts available outside the affected application. Identify who can approve a temporary workaround and what business work can continue.
Follow through on findings
Check security advisories against the installed versions and configuration. Assign remediation and verify completion. For suspected compromise, preserve evidence and use the incident process.
With the responsible leads, write down the privacy, contractual and sector obligations that apply. Don’t assume GDPR, HIPAA or a US defence framework applies to every Canadian business. Revisit the dependency map when software or suppliers change.


