Load an SBOM

Drop your SBOM here CycloneDX JSON / XML · SPDX JSON · up to 25 MB

The SBOM is parsed in your browser and never uploaded. For vulnerability matching only package identifiers (name, version, PURL, CPE) are sent to the NetTools server, which queries NVD; the scan is held in memory for at most an hour.

Component

How it works

The SBOM is parsed in your browser (CycloneDX 1.2–1.6 JSON/XML, SPDX 2.2/2.3 JSON): components, metadata, licenses and dependency relationships, from which direct and transitive dependencies, depth and dependency paths are derived. For vulnerability matching only package identifiers — name, version, PURL, CPE — are sent to the NetTools server, which resolves each package to NVD CPE names (via the CPE dictionary or the CPE in the SBOM) and asks NVD which CVEs apply to that exact version. Results reuse the CVE Search data layer, so every CVE links to its full record.

From SBOM to CVE

  • Parsing (in your browser): CycloneDX 1.2–1.6 JSON/XML and SPDX 2.2/2.3 JSON — metadata, components, licenses, suppliers, PURLs, CPEs and dependency relationships. Direct dependencies are those of the described application; everything reached through them is transitive.
  • Matching (NetTools server → NVD): a CPE in the SBOM is used directly (exact match). Otherwise the PURL name and version are looked up in NVD’s CPE dictionary, using the ecosystem to pick the right vendor (strong match) — or marked as a possible match when that is ambiguous. NVD then decides which CVEs apply to that exact version.
  • Never by name alone: components without a version, a PURL or CPE, or without any NVD CPE are “Unable to determine” instead of guessed.
  • Fixed versions come from NVD’s version ranges (versionEndExcluding) for the matched product only.
  • Risk score: Critical 10, High 5, Medium 2, Low 0.5, unscored 1 point per finding; ×1.5 for direct dependencies; +5 for CISA KEV; possible matches count half; score = 100 × (1 − e^(−points/50)). A NetTools assessment, not an industry standard.

Safeguards & responsible use

  • The SBOM itself never leaves your browser; only package identifiers are sent, and scan jobs are kept in server memory for at most an hour.
  • Packages are never matched by name alone; every match shows its confidence (exact, strong, possible, unable).
  • Fixed versions come only from NVD version ranges; nothing is guessed.
  • An SBOM is data: nothing in it is executed, fetched or resolved.

Frequently asked questions

Which SBOM formats are supported?

CycloneDX 1.2–1.6 in JSON or XML, and SPDX 2.2 / 2.3 in JSON. SPDX tag-value, SPDX RDF/XML and SPDX 3.0 are not supported yet.

Is my SBOM uploaded?

No. It is parsed in your browser. Only package identifiers (name, version, PURL, CPE) are sent to the NetTools server for the NVD lookup, and that scan job is deleted after an hour or when you press Clear.

Why is a component “Unable to determine”?

It has no version, no PURL or CPE, or NVD has no CPE for that package. Matching it by name alone could report CVEs of a different product, so NetTools does not guess.

Why does a scan take several minutes?

Each unique package needs NVD lookups, which are paced within NVD’s public rate limits. Results are cached, so scanning the same packages again is much faster.

How is the risk score calculated?

Per finding: Critical 10, High 5, Medium 2, Low 0.5, unscored 1 point; ×1.5 for direct dependencies; +5 for CVEs in CISA KEV; possible matches count half. Score = 100 × (1 − e^(−points/50)). It is a NetTools assessment, not an industry standard.