Why Cull?

Type a product the way people actually write it: Apache 2.4.49, a line of scanner output, a CPE, a package URL. Cull works out which product you mean, which of its advisories apply to exactly that version, and why. We measured both halves against the tools people use today.

Twice as many products found from plain names, and on exact CPEs the strongest version-aware result in this benchmark: the most confirmed findings, the fewest known misses, a tenth of Search Vulns' false positives and the lowest response latency.

Step one: which product is it?

Plain product names in, catalogue products out · measured 2026-10-24

2x the products found
100%50%0% 88.5% 44.7% 62.3% 28.7% 74.5% 66.2% CullSearch CullSearch CullSearch ENDOFLIFE.DATEPRODUCT FOUND ENDOFLIFE.DATEVERIFIED PRODUCT WAPPALYZERVERIFIED PRODUCT
The inputs are the product names endoflife.date and Wappalyzer use, written the way their pages write them. "Verified" means the answer is the product behind the dataset's own CPE or package identifier. That understates Cull: endoflife.date often lists only an empty container image, so a correct answer under the product's real CPE counts as wrong about 40 times. In a stratified random sample of 100 product names, package names and CPEs, Cull identified a product for 98 and Search Vulns for 58.

Step two: which advisories apply?

To compare the matching itself, we took identification off the table and gave every tool the same 50 exact CPE queries: 25 vendor-documented fixes, each queried once on the last affected release and once on the first fixed one. Exactly the place where a range that is off by one release produces a false alarm or a silent miss. Findings are checked against a frozen reference set; vendor and CNA records take precedence over imported NVD data.

Cull, Search Vulns, cve-search and NVD completed all 50 queries. cve-search ran locally and ignores the submitted version, so it appears here as a version-blind candidate-retrieval baseline: what you get when a product lookup is treated as a version answer. Vulners returned 31 answers before its quota failed.

Extended 2026-08-29 · 25 affected/fixed pairs · exact product CPEs

Quality
Each case/CVE pair is marked as a true positive, a false positive or a known miss.
Reference
The expected results come from reviewed vendor, CNA and first-party records frozen for this run.
Speed
Latency includes the full response, body download and JSON decoding. Quota waiting is excluded.
Scope
LOCALHOST identifies the local cve-search run. 31/50 identifies Vulners' partial run.
Can an ordinary user search by product and version?
ToolDirect queryPractical input workflow
CullYesFree text such as Apache 2.4.49, scanner banners, aliases, CPEs, PURLs and package coordinates are accepted directly.
Search VulnsYesAccepts a product query directly; this quality run still used the same exact CPE as the other version-aware tools.
cve-searchPartialRequires vendor and product fields and does not evaluate the submitted version.
VulnersNoThe measured software-audit endpoint requires an exact software coordinate such as a CPE.
NVDTwo stepsSearch the separate CPE dictionary, choose the correct CPE, then submit it to the CVE API. The benchmark supplied that CPE.
The answer-quality charts start after product identification. This table makes the otherwise hidden CPE-discovery burden explicit; it is deliberately not disguised inside a combined score.

Answer quality

50 exact product/version CPEs · Vulners returned 31

Best version-aware result

True positives higher is better

250012500 2383 2377 2177 1408 2146 CullSearchcve-s.VulnersNVD NO VERSION31/50

False positives lower is better · axis break

700065003001500 24 250 6977 237 17 CullSearchcve-s.VulnersNVD NO VERSION31/50

Known misses lower is better

2801400 39 45 245 52 276 CullSearchcve-s.VulnersNVD NO VERSION31/50
Cull returned 2,383 confirmed case/CVE matches, 24 false positives and 39 known misses. Search Vulns found nearly as much but flagged ten times as many findings that do not apply; NVD flagged even fewer, 17, but missed seven times as many. Uncertain Cull review candidates are not promoted to affected findings. The cve-search figures show what a product-only lookup looks like when treated as a version answer; its false-positive chart skips the unused range between 300 and 6,500. Vulners covers only 31 queries.

Recall

Share of known relevant findings returned

100%50%0% 98.39%Cull 98.14%Search 89.88%cve-s. 96.44%Vulners 88.60%NVD NO VERSION31/50
Recall shows how much of the reviewed affected reference set each provider returned. Higher is better.

Complete-response latency

Milliseconds · quota pacing excluded

p50p95
10505000 ms 18.1 / 29.1Cull 33.0 / 120.4Search 136.8 / 1010.7cve-s. 56.8 / 148.9Vulners 560.9 / 835.7NVD NO VERSION · LOCALHOST31/50
Cull's p95 was 29.1 ms, faster than the median of every other tool; the slowest measured p95 was local cve-search at 1,010.7 ms.
Limits of the benchmark. A “known miss” is a CVE in this frozen reference set that a tool did not return. The counts are case/CVE pairs, not unique CVEs across the whole dataset. These 50 queries are useful regression cases, but they are not a market-weighted sample of all software. cve-search is included to show the output of a version-blind product lookup. Vulners produced 31 valid answers and is not extrapolated to 50.

All 50 benchmark queries

The exact CPE strings used for the run.

  1. OpenSSL 1.0.1fcpe:2.3:a:openssl:openssl:1.0.1f:*:*:*:*:*:*:*
  2. OpenSSL 1.0.1gcpe:2.3:a:openssl:openssl:1.0.1g:*:*:*:*:*:*:*
  3. nginx 1.20.0cpe:2.3:a:f5:nginx:1.20.0:*:*:*:*:*:*:*
  4. nginx 1.20.1cpe:2.3:a:f5:nginx:1.20.1:*:*:*:*:*:*:*
  5. Apache Log4j 2.14.1cpe:2.3:a:apache:log4j:2.14.1:*:*:*:*:*:*:*
  6. Apache Log4j 2.17.1cpe:2.3:a:apache:log4j:2.17.1:*:*:*:*:*:*:*
  7. Apache HTTP Server 2.4.49cpe:2.3:a:apache:http_server:2.4.49:*:*:*:*:*:*:*
  8. Apache HTTP Server 2.4.51cpe:2.3:a:apache:http_server:2.4.51:*:*:*:*:*:*:*
  9. curl 8.3.0cpe:2.3:a:haxx:libcurl:8.3.0:*:*:*:*:*:*:*
  10. curl 8.4.0cpe:2.3:a:haxx:libcurl:8.4.0:*:*:*:*:*:*:*
  11. Spring Framework 5.3.17cpe:2.3:a:vmware:spring_framework:5.3.17:*:*:*:*:*:*:*
  12. Spring Framework 5.3.18cpe:2.3:a:vmware:spring_framework:5.3.18:*:*:*:*:*:*:*
  13. OpenSSH 9.7p1cpe:2.3:a:openbsd:openssh:9.7:p1:*:*:*:*:*:*
  14. OpenSSH 9.8p1cpe:2.3:a:openbsd:openssh:9.8:p1:*:*:*:*:*:*
  15. sudo 1.9.5p1cpe:2.3:a:sudo_project:sudo:1.9.5:patch1:*:*:*:*:*:*
  16. sudo 1.9.5p2cpe:2.3:a:sudo_project:sudo:1.9.5:patch2:*:*:*:*:*:*
  17. XZ Utils 5.6.1cpe:2.3:a:tukaani:xz:5.6.1:*:*:*:*:*:*:*
  18. XZ Utils 5.6.2cpe:2.3:a:tukaani:xz:5.6.2:*:*:*:*:*:*:*
  19. runc 1.1.11cpe:2.3:a:linuxfoundation:runc:1.1.11:*:*:*:*:*:*:*
  20. runc 1.1.12cpe:2.3:a:linuxfoundation:runc:1.1.12:*:*:*:*:*:*:*
  21. libwebp 1.3.1cpe:2.3:a:webmproject:libwebp:1.3.1:-:*:*:*:*:*:*
  22. libwebp 1.3.2cpe:2.3:a:webmproject:libwebp:1.3.2:*:*:*:*:*:*:*
  23. Atlassian Confluence Server 8.5.1cpe:2.3:a:atlassian:confluence_server:8.5.1:*:*:*:*:*:*:*
  24. Atlassian Confluence Server 8.5.2cpe:2.3:a:atlassian:confluence_server:8.5.2:*:*:*:*:*:*:*
  25. GitLab Enterprise Edition 16.7.1cpe:2.3:a:gitlab:gitlab:16.7.1:*:*:*:enterprise:*:*:*
  26. GitLab Enterprise Edition 16.7.2cpe:2.3:a:gitlab:gitlab:16.7.2:*:*:*:enterprise:*:*:*
  27. Jenkins 2.441cpe:2.3:a:jenkins:jenkins:2.441:*:*:*:-:*:*:*
  28. Jenkins 2.442cpe:2.3:a:jenkins:jenkins:2.442:*:*:*:-:*:*:*
  29. JetBrains TeamCity 2023.11.3cpe:2.3:a:jetbrains:teamcity:2023.11.3:*:*:*:*:*:*:*
  30. JetBrains TeamCity 2023.11.4cpe:2.3:a:jetbrains:teamcity:2023.11.4:*:*:*:*:*:*:*
  31. Fortinet FortiOS 7.2.4cpe:2.3:o:fortinet:fortios:7.2.4:*:*:*:*:*:*:*
  32. Fortinet FortiOS 7.2.5cpe:2.3:o:fortinet:fortios:7.2.5:*:*:*:*:*:*:*
  33. Palo Alto Networks PAN-OS 10.2.9cpe:2.3:o:paloaltonetworks:pan-os:10.2.9:-:*:*:*:*:*:*
  34. Palo Alto Networks PAN-OS 10.2.9-h1cpe:2.3:o:paloaltonetworks:pan-os:10.2.9:h1:*:*:*:*:*:*
  35. ConnectWise ScreenConnect 23.8.5cpe:2.3:a:connectwise:screenconnect:23.8.5:*:*:*:*:*:*:*
  36. ConnectWise ScreenConnect 23.9.8cpe:2.3:a:connectwise:screenconnect:23.9.8:*:*:*:*:*:*:*
  37. Fortra GoAnywhere MFT 7.4.0cpe:2.3:a:fortra:goanywhere_managed_file_transfer:7.4.0:*:*:*:*:*:*:*
  38. Fortra GoAnywhere MFT 7.4.1cpe:2.3:a:fortra:goanywhere_managed_file_transfer:7.4.1:*:*:*:*:*:*:*
  39. Progress WS_FTP Server 8.8.1cpe:2.3:a:progress:ws_ftp_server:8.8.1:*:*:*:*:*:*:*
  40. Progress WS_FTP Server 8.8.2cpe:2.3:a:progress:ws_ftp_server:8.8.2:*:*:*:*:*:*:*
  41. Joomla 4.2.7cpe:2.3:a:joomla:joomla\!:4.2.7:*:*:*:*:*:*:*
  42. Joomla 4.2.8cpe:2.3:a:joomla:joomla\!:4.2.8:*:*:*:*:*:*:*
  43. Grafana 8.3.0cpe:2.3:a:grafana:grafana:8.3.0:*:*:*:*:*:*:*
  44. Grafana 8.3.1cpe:2.3:a:grafana:grafana:8.3.1:*:*:*:*:*:*:*
  45. Apache Tomcat 9.0.30cpe:2.3:a:apache:tomcat:9.0.30:*:*:*:*:*:*:*
  46. Apache Tomcat 9.0.31cpe:2.3:a:apache:tomcat:9.0.31:*:*:*:*:*:*:*
  47. Git 2.45.0cpe:2.3:a:git:git:2.45.0:*:*:*:*:*:*:*
  48. Git 2.45.1cpe:2.3:a:git:git:2.45.1:*:*:*:*:*:*:*
  49. PostgreSQL 16.4cpe:2.3:a:postgresql:postgresql:16.4:*:*:*:*:*:*:*
  50. PostgreSQL 16.5cpe:2.3:a:postgresql:postgresql:16.5:*:*:*:*:*:*:*
How it works

How a search becomes a verdict

Vulnerability data looks simple until you try to answer one question with it: is this exact thing vulnerable? The answer hides in a dozen feeds that name the same product five different ways, number releases by different rules, and regularly contradict each other. This page walks through what Cull does about that, with the traps we fell into on the way. Most of them you would never think of until a wrong answer showed you.

2,172,616advisories in memory
612,772resolved products
25independent sources
4explicit API verdict states

In short: collect the records, keep what each source actually said, work out which product the query means, compare the version by the rules of the ecosystem it came from, and return the evidence with the answer. Never turn "we could not tell" into "you are fine".

CollectDownload advisory, package and exploit data
NormalizeParse each format without dropping its scope
ConnectLink names, CPEs, package URLs and advisory ids
DecideCompare the requested version with the evidence
UseReview the result in the browser or call the API
01 · Sources

Where the data comes from

No single feed knows everything, and they disagree more often than you would hope. NVD describes upstream ranges, a Linux distribution knows which fixes it backported, Metasploit knows what actually fires. Cull reads 25 sources and keeps track of what each one is entitled to say.

Core records

Published vulnerabilities

NVD, the CVE List and EUVD supply identifiers, descriptions, CPE configurations, severity and CNA product statements.

Fixes

Vendor and distribution advisories

curl, Red Hat, Debian, Ubuntu, Alpine, Wolfi, Chainguard, AlmaLinux and Oracle Linux provide package fixes and backport status; SUSE arrives through OSV.

Packages

Ecosystem advisories

OSV and GitHub connect vulnerabilities to package URLs and native version ranges for npm, PyPI, Go, crates, Maven, RubyGems, NuGet, GitHub Actions and other registries. Bitnami adds upstream ranges for the server applications it packages.

Exploitation

Priority and tooling

EPSS, CISA KEV, SSVC, Exploit-DB, Metasploit, Nuclei, PoC-in-GitHub and Wordfence add likelihood, observed exploitation and available tooling.

Lifecycle

Supported release lines

endoflife.date adds upstream active-support and end-of-life dates for exact product versions. Cull keeps this operational context separate from CVE applicability.

Older is not always worse

Debian's curl 7.88.1 has fixes upstream shipped in 8.x

Compare 7.88.1 with upstream's "fixed in 8.x" and you report holes Debian closed months ago. 7.88.1-11+deb12u10 carries the backports. When the query names Debian or the version gives the package away, Debian's own statement decides.

An exploit is not a verdict

A Metasploit module proves tooling, not exposure

The module tells you exploitation is cheap. Whether your package is affected is still decided by the version evidence. Both facts stay on the finding, neither overwrites the other.

Two kinds of end of life

Upstream EOL is not your distribution's EOL

A release upstream abandoned can be fully supported in an enterprise distribution for years. Cull never substitutes one date for the other.

Configuration matters

Some flaws need a feature switched on

A NetScaler flaw that only applies to appliances configured for SAML is shown with that prerequisite and its Citrix source, and with the honest note that Cull has not checked your configuration.

What is kept

Native identifiers, aliases, range types, fix statements, withdrawn status, scores, references and exploit provenance stay attached to their source. They are still available when Cull produces a verdict and when someone reviews it.

02 · Import

Keeping the meaning of each source

Every feed uses a different schema, and the dangerous part of an import is not the file that fails to parse. It is the file that parses fine and quietly means something else. The importers translate each schema into Cull's model while keeping the source, the exact range boundaries and the difference between affected, fixed, not affected, withdrawn and unknown.

  • 1Use a parser for each source. CPE configurations, OSV events, distro ranges, aliases and exploit references do not pass through a generic text importer.
  • 2Connect aliases without erasing them. A CVE and a source-native advisory id can refer to the same issue while remaining traceable to their original records.
  • 3Keep range boundaries exact. Inclusive, exclusive, exact, introduced, last-affected, fixed and open-ended ranges remain different cases.
  • 4Stop a bad import. Source volume bounds, parser skip ratios, corpus shrinkage checks and identity fan-out limits block incomplete or implausible datasets.
Same fix, different statement

Is 1.4.3-rc1 affected?

Under "fixed in 1.4.3" it is: a release candidate sorts before the release. Under "last affected 1.4.2" it is not. Flatten both into "before 1.4.3", as many importers do, and one of them is wrong. Cull keeps them apart.

A quiet catastrophe

The feed that suddenly halves

A format change that makes a parser skip half the records does not crash anything. It just publishes half as many vulnerabilities, which looks exactly like good news. Volume bounds and skip ratios make that build fail instead.

Not every range is a range

Commit hashes are not versions

OSV sometimes bounds a range with git commits. Cull cannot compare a release number with a hash, so it marks the statement unverifiable instead of guessing, and the finding says why it could not decide.

The letter that sorts backwards

OpenSSL 1.0.2u is not 1.0.2-u

NVD files it as version 1.0.2 with update u. Join them with a dash and 1.0.2-u becomes a pre-release that sorts below 1.0.2. Cull attaches a single letter directly, so the patch level sorts above the release it patches.

source Aaffected before 2.4.50
source Bfixed in package 2.4.49-3+deb11u1
resulttwo scoped statements, not one overwritten field
03 · Identity

One product, many names

This is where most vulnerability tools quietly lose findings. The same software appears as an NVD CPE, an SBOM package URL, a distribution package, an EUVD label and two former company names. Every one of those keys holds part of the product's history. Search one of them and the advisories filed under the others are simply invisible, and the result looks clean.

Split in half

Checkmk: 108 advisories here, 10 over there

NVD files projects without a company under check_mk_project and later under the company's own name checkmk. Varnish once reported not affected for 6.0.0 from the three-record spelling while Varnish Cache had eleven advisories for that release.

Acquisitions leave fossils

VirtualBox 4.3.40: 5 records or 245?

Sun's old sun:virtualbox entry holds five advisories; Oracle VM VirtualBox holds 245 for that release. Ask the old name and you get an all-clear. Lotus Domino lives on as ibm:lotus_domino, JBoss as Red Hat's.

Cull therefore groups identifiers into one product entity while building the dataset. The links come from ten places, each with its own evidence:

  • 1Separator variants. sd_675_firmware and sd675_firmware are linked. A changed letter or meaningful symbol is not ignored: 865 and 865+ are different chips.
  • 2Vendor spellings. palo_alto_networks and paloaltonetworks, gpac and gpac_project, kaspersky and kaspersky_lab: words that only describe the organisation do not make another vendor. The product itself still has to match.
  • 3A brand inside the product name. After an acquisition the new owner often keeps the old brand in front: lotus:domino is ibm:lotus_domino. A generic two-word name that many vendors carry, such as web_server, never qualifies.
  • 4NVD renames. Cull uses NVD's deprecatedBy field when it describes a clean one-to-one rename. A key that splits into several successors is left alone.
  • 5Registry links. If a CPE points to npmjs.com/package/lodash, that is direct evidence that it names the npm package.
  • 6Package and CPE with the same releases. A registry package named like a CPE product is that product when both share most of their CVEs and every comparable shared CVE names the same fixed release. That joins WooCommerce, Wagtail or Gogs with their CPEs, about 5,600 pairs in all.
  • 7Overlap between schemes. An opaque EUVD product key is linked to a CPE only when an exact name and shared CVEs support the match. Generic names such as “Reader” are not enough.
  • 8Reviewed pairs. About 210 acquisition and rebrand links cover cases rules cannot infer safely, including Sun to Oracle and Macromedia to Adobe. Each entry must state its evidence.
  • 9Reviewed names. Familiar spellings that registries do not provide: iOS for NVD's iphone_os, k8s for Kubernetes, MSSQL for SQL Server.
  • 10Component relations. A suite can point to one of its parts without becoming the same product. Versions cross that relation only where the comparison is safe.
Same name, same CVEs, different product

The Apache Avro Rust crate is not Apache Avro

It shares the name and some CVEs, but numbers its releases 0.x while the Java line is at 1.11. Merge them and every Rust release becomes "affected" by every Java bug below 1.11. Requiring the same fixed releases keeps bindings, bundled copies and PubNub's per-language SDKs apart.

One theme, two vendors

When a package claims two products, it links neither

The WordPress theme ask-me carries the CVEs of two different vendors' "Ask Me" products. Linking both through it would quietly merge them, so it links nothing. Reviewed controls such as curl versus libcurl, Firefox versus Firefox ESR and FileZilla client versus server must stay apart, or the build does not ship.

A union-find pass merges the accepted links so searches do not have to walk the graph. No product may collect more than 64 identifiers. If a bad rule crosses that limit, the build fails instead of publishing an oversized product group.

Several identifiers merging into one product entity cpe .../http_server cpe .../http-server cpe redhat/httpd purl rpm/httpd euvd:9f2c... linked by spelling & renames registry refs & CVE overlap one product Apache HTTP Server union-find, capped at 64 names
Spelling rules, renames, registry links, shared releases and reviewed mappings can all point to the same product. A group larger than 64 identifiers causes the build to fail.
A shared CVE alone is not enough

Bundles and dependencies often appear in the same advisory: Chrome and the Skia library it contains share CVEs and stay separate. Shared CVEs only count together with a matching name and matching fixed releases, and reviewed exclusions keep known mixed EUVD keys apart.

05 · Version comparison

Comparing versions

Finding the product is half the job. Cull must then place the submitted version inside or outside every relevant advisory range, and "is 1.2 before 1.3" turns out to have about a dozen answers. It parses the version into segments and chooses comparison rules from the package ecosystem or distribution it came from.

1.2rc1 sorts before 1.2, while 1.2p1 sorts after it. The same string can change meaning between ecosystems: SemVer treats 1.0.0-1 as a pre-release, while Debian treats the suffix as a package revision. Maven, RubyGems, NuGet and Alpine have their own ordering rules as well.

The version ordering ladder and the epoch trap pre-release release patch and update newer alpha beta rc 1.2 1.2p1 1.2u3 the epoch trap rpm -q prints 2.4.37-43.el8_5, no epoch When input omits the epoch, cull compares the visible version and release instead.
Pre-releases sort below the release and patches above it. Package epochs need special handling because command output often leaves them out.
The invisible digit

The epoch trap

Red Hat ships Apache with epoch 1, written 1:2.4.37, but rpm -q may print only 2.4.37-43.el8_5. Read that as epoch 0 and a fixed server looks vulnerable to everything since 2018. If only one side has an epoch, Cull compares the visible version and release; if both do, normal RPM ordering applies.

Hotfixes and patch levels

PAN-OS 10.2.9 and 10.2.9-h1 are different answers

So are OpenSSH 9.7p1 and 9.8p1, and sudo 1.9.5p1 and 1.9.5p2. NVD stores the suffix in a separate CPE field; Cull folds it back so the release you type is the release that is compared. All three pairs are in the benchmark for exactly that reason.

Same product, different editions

GitLab Enterprise is not GitLab Community

Both share one CPE and differ only in its edition field. A statement scoped to Enterprise must not confirm, or clear, a Community query, so the edition is kept as context rather than folded into the product.

Backports need a home

A plain apache 2.4.49 does not borrow Debian's fixes

Distribution rules apply only when the query names a distribution or contains a recognizable package version such as 7.88.1-10+deb12u5. If the only evidence belongs to some other distribution, Cull leaves the result undecided rather than picking one.

06 · Verdicts

Affected, ruled out or still unknown

Each relevant advisory receives a result and a reason. The most important design choice on this page: unknown stays separate from ruled out, so a lack of usable range data can never look like a clean result.

Verdicts sorted into affected, ruled out, and could not check advisory vs your version Affected in range, act on it Ruled out out of range, or fixed Could not check exists, but nobody said distro says fixed clears it, keeps the split only fixed, not affected or out of range can clear
A distribution backport can clear an upstream match. Both source statements remain attached to the result, and missing evidence stays in its own bucket.
  • AAffected. The version is inside an affected range or an advisory names it directly.
  • BRuled out. The version is outside the affected ranges, fixed, or explicitly marked not affected.
  • CCould not check. The advisory is relevant, but its version or distribution scope cannot answer this query.
The empty array

Nothing found is not the same as nothing there

In automation an empty result looks like success. So an unknown advisory id is reported as unchecked instead of dropped, a product Cull could not identify is UNRESOLVED rather than "not affected", and an uncertain product match can only ever produce "unknown".

Clearing needs proof

A withdrawn record cannot clear anything

A distribution may overrule an upstream match only with positive evidence: fixed, not affected or out of range. An inconclusive or withdrawn record cannot, and a finding is only withdrawn when every source that reported it has retracted.

The API exposes the same distinction as AFFECTED, NOT_AFFECTED, UNKNOWN_OR_UNCHECKED and UNRESOLVED.

07 · Exploitation

Exploit evidence and urgency

An affected result does not say whether anyone can actually exploit it this afternoon. Cull keeps exploit evidence separate from CVSS and shows where that evidence came from.

EPSS estimates likelihood. CISA KEV records observed exploitation. SSVC provides decision context. Exploit-DB, Metasploit, Nuclei and PoC-in-GitHub indicate what public tooling exists. Wordfence adds coverage for WordPress plugins that often have no CVE at all.

The strongest evidence determines the finding's exploit tier:

The five-rung exploit ladder unreviewed proof of concept verifiable weaponized observed PoC in GitHub Exploit-DB Nuclei template Metasploit module CISA KEV, in the wild
The scale runs from unreviewed code to observed exploitation. Verified tool identifiers can also be used to build a command.

Results are ordered by this evidence before severity: known-exploited first, then any exploit, then EPSS probability, and only then CVSS. A 9.8 that nobody can reach sits below a 7.5 with a maintained Metasploit module. Severity, EPSS, publication date and identifier remain available as explicit sort keys, and the tier itself is a property of the finding that does not change with the order. If Cull has a validated Nuclei template id, Metasploit module path or Exploit-DB id, it can show the corresponding command. A reference URL is never inserted into a shell command: a link in an advisory is somebody else's text.

Scores are not interchangeable

CVSS describes technical severity, EPSS estimates likelihood, KEV records observed exploitation and SSVC adds decision context. If one of those values is missing, Cull leaves it missing rather than filling the gap with another score.

08 · Research workflow

Working with the results

Cull's research workbench keeps the finding list next to the selected CVE. Applicability, source statements, aliases, affected ranges, fixes, exploit tooling and the event timeline stay in one place, so you never need five tabs to answer one question.

Filter

Reduce the result set

Search ids, descriptions, CWE, products, sources and tooling. Structured filters cover CVSS, EPSS, KEV, exploit type, certainty, fixes, dates, vector fields and attack primitives.

Inspect

Read the evidence

The inspector shows source verdicts, disagreements, remediation, affected branches, commands and the timeline for the selected finding.

Queue

Keep a working list

Save findings locally in the browser, add comments, reorder them and filter the queue by KEV, tooling or severity.

Export

Hand the work over

Copy selected findings as Markdown, CSV or JSON. Queue exports include notes, and verified Nuclei templates can be combined into one command.

The query language and the controls use the same fields. openssl 1.0.1 is:kev cvss:>=9 vector:AV:N can be bookmarked and rerun. Filters are applied before totals, sorting and pagination, so the displayed count matches the filtered result.

No combined priority score

CVSS, EPSS, known exploitation, exploit maturity, fix availability and publication date remain separate fields. You can choose which one matters for the current job.

09 · API and automation

Using the API

The HTTP API calls the same matcher as the web interface. It accepts the same targets and filters, returns the same verdict states and adds pagination and stable response schemas for integrations.

GET /api/v1/search

Search one product, package coordinate or advisory id. Results can be filtered and sorted, or reduced to ids with view=ids.

POST /api/v1/bulk

Check up to 100 targets in JSON or newline-delimited text. Compact output is available for inventories and boundary checks.

GET /api/v1/openapi.json

Download the OpenAPI 3.1 description of parameters, result states, findings, pagination and errors.

The full response includes aliases, source verdicts, affected and ruled-out evidence, unchecked records, CVSS, EPSS, KEV, SSVC, CWE, attack primitives, exploit references, commands, timelines, fixes and source disagreements. CI jobs that only need identifiers can request view=ids.

The status field removes ambiguity around empty results. AFFECTED has applicable evidence, NOT_AFFECTED resolved the target without one, UNKNOWN_OR_UNCHECKED has relevant but inconclusive evidence, and UNRESOLVED could not identify the product. A product chosen without decisive evidence carries product_uncertain, so an inventory tool can ask a human before it stores the mapping.

One matcher

The workbench, single-search endpoint and bulk endpoint render the same underlying result. Browser and API behavior therefore change together.

10 · Serving engine

How the dataset fits in memory

Cull serves about 2,172,616 advisories from immutable in-memory structures on one machine, which is why a complete answer arrives in about 18 milliseconds. In the measured 2026-08-24 corpus, the SQLite database was 2.4 GB and the separate prose file 558 MB. The loaded server used about 1.6 GiB of live Go heap and 2.5 GiB steady RSS.

Structured matching data lives in the Go heap. Titles, descriptions and reference URLs live in a separate memory-mapped file because the matcher does not need them while deciding applicability. The file contains no Go pointers, so the garbage collector does not scan it, and the operating system can discard and reload those pages under memory pressure.

Repeated values are interned. In that corpus, 12.8 million range-list occurrences reduced to 311 thousand distinct values, and 3.1 million parsed-version occurrences reduced to 250 thousand. Parsed versions use one allocation each. Temporary build maps are discarded before the server reports ready.

GOMEMLIMIT is not compression

GOMEMLIMIT cannot shrink live data. A value below the working set only makes garbage collection more frequent. The actual savings come from compact structures, interning and the pointer-free mapped file.

11 · Evaluation and improvement

How mistakes become tests

Every example on this page was once a wrong answer. Most matching bugs live at boundaries: a rename, an exclusive upper range, a package epoch, a backport, an alias collision, a scanner that splits a version in two. Once one is fixed, it stays in the regression suite, so it cannot come back.

Refused for being too exact

mysql 8.0.30

A distribution package named exactly "mysql 8.0" won the version-bearing reading, was then set aside as packaging, and the query was refused although "mysql" matched perfectly. Now every reading is kept and the next one answers.

Too much catalogue

haproxy 2.6.0 and laravel 9.0.0

Both stopped resolving once the catalogue held a source-local entry of exactly that word. The vendor's own product named after itself now wins, while "adobe" still asks.

The wrong OpenSSL

OpenSSL 1.1.1w

The release branch in the name found SUSE's parallel package openssl-1_1. A release branch read into a name can no longer pick distribution packaging the query did not ask for.

A distribution called Linux

Alpine Linux 3.23.7

Read word by word it became Alpine context around a product called "linux". Distribution names are now read as whole phrases, and a full name such as Fedora Linux stays a product.

Unit tests

Keep the awkward examples

Hundreds of focused automated tests cover version rules, query parsing, product links, source precedence, verdicts, filters, pagination and API schemas.

Integration

Check the whole answer

Corpus fixtures run through import, matching and response rendering. Tests check the bucket, reason, evidence, aliases, fixes and status around an id.

Data checks

Reject suspicious refreshes

Freshness, source volume, parse coverage, corpus shrinkage and product fan-out are checked before a dataset can serve searches, and reviewed identity controls must hold.

Benchmark

Rerun the reviewed cases

The 50-query benchmark uses fixed inputs and a reviewed reference set, and thousands of real product names from scanners and catalogues test identification. Both record misses separately from false alarms.