Forty Findings, One Afternoon
Everything is 'high'. Decide what actually gets fixed first.
Your scanner produced forty findings and labelled most of them 'high'. You have one afternoon and one engineer. This exercise is about the step nobody teaches: turning a severity-sorted dump into a defensible priority order using exposure, exploitation evidence, and business context โ the approach set out in Integrating Threat Intelligence. Read the scanner output, the exploitation feed, and the asset inventory together, because none of them is sufficient alone.
Fictional training scenario. All output is synthetic. CVE identifiers shown are illustrative placeholders and do not refer to real advisories โ the exercise is about the prioritization method, not about looking up specific vulnerabilities.
Your brief
You're the security engineer at Northwind. A nuclei scan of your external estate finished overnight, you have this week's exploitation feed, and you have the asset inventory that says what each host actually does. Your manager wants to know what gets fixed today. Read all three and decide.
critical Remote code execution in FreightRouter (CVE-2026-11001) https://legacy-router.northwind.example high Outdated TLS configuration โ TLS 1.0 enabled https://marketing.northwind.example high Exposed .git directory https://api.northwind.example high Authenticated RCE in AdminPanel (CVE-2026-11042) https://admin.northwind.internal high Default credentials accepted https://printer-3f.northwind.internal high SQL injection in order lookup (CVE-2026-11077) https://shop.northwind.example medium Missing security headers https://www.northwind.example medium Directory listing enabled https://cdn.northwind.example medium Verbose error messages https://blog.northwind.example
Triage decisions5 questions ยท 70% to pass
0/5 answered
Defensive lessons
- 1Severity is an input, not an answer. CVSS describes a vulnerability in the abstract; your risk is exposure ร confirmed exploitation ร business impact. Sorting by severity alone produces a queue nobody can defend.
- 2Confirmed in-the-wild exploitation is the highest-value filter available, and it's free. Anything in your estate that appears on an exploitation feed goes to the top regardless of its severity label.
- 3Cross-reference findings against an asset inventory before acting. A 'critical' on a decommissioned host and a 'high' on your payment platform are not comparable, and only the inventory tells you which is which.
- 4A dangling DNS record is its own finding. Decommissioning isn't finished until the DNS record is gone, or someone else can claim the name and serve content from your domain.
- 5'Internal-only' lowers exposure without granting immunity. Rank internal findings below internet-facing ones, but keep admin consoles and RCE high โ they're what an attacker reaches for after the first foothold.
- 6Exposed .git directories are credential leaks, not configuration nits. Source, config, and commit history frequently contain keys that outlive the code.
- 7Write the prioritization rule down. A documented, tunable formula survives the 'why is this ranked above that?' conversation. Intuition does not, and it can't be handed to the next person on call.