Every term this app puts on a chart, in plain language — what it means, why it matters, and how it shows up here. Open “More” for the long version with examples.
A public, unique ID for one specific security flaw in one specific product, so everyone talks about the same issue.
The *type* of coding mistake behind a vulnerability, e.g. 'SQL injection'. Many CVEs share one CWE.
A 0–10 technical severity score for how bad a flaw is if exploited. It measures damage potential, not likelihood.
A 0–100% probability that a vulnerability will be exploited in the wild in the next 30 days.
A government-maintained list of vulnerabilities confirmed to be exploited in real attacks right now.
One run of the collection agent against either a single machine (PC scan) or a range of addresses (network scan).
One machine or network address that a scan looked at.
One observation from a scan — a package, service or open port — optionally paired with a candidate CVE.
The three-part identity used to line up something you run with the software a CVE affects.
How strongly the app believes a CVE really applies to something a scan found, from 0% to 100%.
Where the vulnerability intelligence in this app comes from: NVD (NIST), MITRE CVE, OpenCVE, CISA KEV and EPSS (FIRST.org).
How many CVEs were published in each time bucket, with the bucket size chosen to match your selected range.
Something is listening or a name was guessed from a port number — never proof of a product or a vulnerability.
An observation that could not be resolved to a trustworthy product identity and must be confirmed before action.
A concrete product and version were read directly from the asset or from the service itself.
A named product was observed, but without a reliable version.
A successful TCP handshake against a host and port. Stored as weak evidence.
The scan did not cover its whole intended scope, so gaps are unknown rather than clean.
Nothing responded, so the run proves nothing about the target.
Hosts responded, but nothing crossed the extraction thresholds.
Collection succeeded on the machine, but no observations reached the server.
The server refused the upload because the payload was malformed or the agent is out of date.
A CVE the matcher generated for an observation. Not a confirmed exposure.
A candidate with enough evidence for an analyst to review.
A high-confidence finding on a real product identity — the only kind counted in patch and SLA metrics.
How recently a host or scan type was successfully covered.
How much we trust the mapping from a raw product name to a normalized vendor/product/version.