Evidence comparison
Semgrep and SonarQube: what the public record shows about each
Both products turn up on devsecops and application security shortlists, so this page puts their two evidence records next to each other: 24 items on Semgrep and 32 on SonarQube, each one linked to the source it came from. It does not name a winner, and the section below explains why the numbers cannot be subtracted from one another.
Read the counts correctly
More advisories does not mean less secure. A vulnerability register records what researchers found and what the vendor disclosed. A product deployed in thousands of organisations with a mature disclosure process accumulates far more entries than a product nobody audits. A low count can mean a small install base, a young product or a vendor that publishes little, and a high count often marks the opposite.
The two counts are not measured with equal precision. An advisory is only listed against a vendor when the register ties it to one of its products by name, and vendors differ enormously in how consistently their products are named. Here that gap is 3 unattributable entries for Semgrep against 16 for SonarQube, which moves the advisory row far more than the products do. Both numbers are in the table so the difference is visible rather than hidden.
Treat this page as two records shown together, never as one score minus another.
The two records side by side
| Measure | Semgrep | SonarQube |
|---|---|---|
| Evidence items in the recordSize of the record, not a quality signal. | 24 | 32 |
| Kinds of source | 3 | 5 |
| Advisories attributed to the vendor's productsCounts scrutiny and deployment breadth. See the caution above. | 0 | 6 |
| Register entries that could not be attributedEntries naming the vendor that could not be tied to one of its products, so they are excluded from the row above. Read the two rows together: where this number is high, the registers describe that vendor's products less consistently, and the advisory count above says more about naming than about the product. | 3 | 16 |
| Of those, in the CISA known-exploited catalogueExploitation observed in the wild. The one count worth reading closely. | 0 | 0 |
| Share from government registers | 13% | 78% |
| Share the vendor cannot edit | 100% | 100% |
| Public practitioner threadsReflects community size as much as product behaviour. | 9 | 0 |
| Most recent entry | Jul 18, 2026 | Jul 18, 2026 |
| Sources last re-queried | Jul 18, 2026 | Jul 18, 2026 |
Where each record comes from
The registers and platforms behind each column, with the body that operates each one. A record resting on government registers carries different weight from one resting on the vendor's own documentation, which is why this is shown per vendor rather than merged.
Semgrep
| Source | Operated by | Classification | Items | Latest |
|---|---|---|---|---|
| European Vulnerability Database | ENISA, European Union | Authoritative | 3 | Apr 30, 2026 |
| Engineering discussions | Hacker News, public threads | Independent | 17 | Jul 09, 2026 |
| Analyst, review and reference sources | Recognised analysts, review marketplaces and technology press | Independent | 4 | Jul 18, 2026 |
SonarQube
| Source | Operated by | Classification | Items | Latest |
|---|---|---|---|---|
| European Vulnerability Database | ENISA, European Union | Authoritative | 16 | Mar 09, 2026 |
| National Vulnerability Database | NIST, United States Department of Commerce | Authoritative | 6 | Mar 09, 2026 |
| CERT-Bund security advisories | BSI, Federal Republic of Germany | Authoritative | 3 | Oct 13, 2025 |
| Analyst, review and reference sources | Recognised analysts, review marketplaces and technology press | Independent | 5 | Jul 18, 2026 |
| Engineering discussions | Hacker News, public threads | Independent | 2 | Feb 03, 2026 |
Documented vulnerabilities
Only entries tied to a product of the named vendor by a structural signal are listed: an advisory reference on the vendor's own domain, the vendor as the assigning authority, or the affected product list naming it. Entries that merely mention a vendor in passing are excluded from both columns.
Semgrep
SonarQube
- CVE-2020-37020highCVSS 7.8Jan 29, 2026
SonarQube 8.3.1 contains an unquoted service path vulnerability that allows local attackers to gain SYSTEM privileges by exploiting the service executable path.
- WID-SEC-2025-2259lowCVSS 4.3Oct 13, 2025
Affected products: SonarSource SonarQube
- CVE-2025-59844highCVSS 7.7Sep 26, 2025
SonarQube Server and Cloud is a static analysis solution for continuous code quality and security inspection.
- CVE-2025-58178highCVSS 7.8Sep 02, 2025
SonarQube Server and Cloud is a static analysis solution for continuous code quality and security inspection.
- WID-SEC-2024-3083mediumCVSS 6.7Oct 07, 2024
Affected products: SonarSource SonarQube
- WID-SEC-2024-1385mediumCVSS 4.9Jun 17, 2024
Affected products: SonarSource SonarQube
What practitioners report
Public threads naming each product in their own title. Not curated for sentiment, and not counted as a verdict: a loud thread is one team's experience, not a measurement.
Semgrep
- Hacker NewsJun 10, 2026Independent
Making Semgrep rip: How Ripgrep inspired us to shave hours off (some) scans
- Hacker NewsMay 28, 2026Independent
We Benchmarked Claude Code, Codex, Semgrep, CodeQL, Trent on 28 CWE-Bench CVEs
- Hacker NewsJul 07, 2025Independent
semgrep: Lightweight static analysis for many languages
- Hacker NewsMar 12, 2025Independent
Upgrading Semgrep from OCaml 4 to OCaml 5
- Hacker NewsFeb 15, 2025Independent
Did Semgrep Just Get a Lot More Interesting?
SonarQube
What neither record can tell you
The two lists below are the honest holes in each record. Everything a choice between these two actually turns on sits outside both of them.
Semgrep
- 3 register entries mention Semgrep, but none of them could be attributed to a product of Semgrep. They describe other vendors' software and are not shown here.
- Pricing, contract terms, notice periods and support commitments are not part of any public register. They come from the vendor and belong in a negotiation record.
- Certification status such as ISO 27001, SOC 2 or BSI C5 has to be checked against the current certificate and its stated scope. A valid certificate can still exclude the product being bought.
SonarQube
- A further 16 register entries mention SonarQube without naming a product of SonarQube as affected. They are excluded rather than counted as vulnerabilities.
- No public practitioner discussion naming SonarQube was found in the indexed threads. Operational experience with this product is not represented in this record.
- Pricing, contract terms, notice periods and support commitments are not part of any public register. They come from the vendor and belong in a negotiation record.
- Certification status such as ISO 27001, SOC 2 or BSI C5 has to be checked against the current certificate and its stated scope. A valid certificate can still exclude the product being bought.
Whether Semgrep is the right choice is not on this page
A record of what is true about a vendor is one half of a decision. The other half is your context: which criteria matter, how heavily each one weighs, what is a hard no, and who has to sign the result off. No public page can know that, and any page that claims to rank vendors for you is guessing at it.
Typical criteria for this category
Needs your case- Detection Accuracy (MITRE ATT&CK coverage)weight
- MDR / Managed Detection & Responseweight
- Deployment Complexityweight
- Automated Response / Remediationweight
- SIEM Integrationweight
Criteria shown as examples for devsecops and application security. Weighting and fit are properties of your decision, not of the vendor, so they are not published here.
What DecisionOS adds on top of this record
- The vendor scored against your criteria with your weighting, not against a generic ranking
- A dealbreaker check against your hard constraints: data residency, contract terms, existing stack
- A side-by-side view of the alternatives on your shortlist, built from this same evidence pool
- This evidence mapped onto your compliance scope: NIS2, DORA, ISO 27001, SOC 2
- A decision memo your board can read, with the trade-offs and the reasoning on the record
Questions this page answers
Which is better, Semgrep or SonarQube?
This page does not answer that, and no public page can. Better exists only relative to a decision: which criteria matter, how heavily each weighs, what counts as a dealbreaker and which systems the product has to live next to. What this page gives you is the evidence on both, 24 items on Semgrep and 32 on SonarQube, each with its source, so the judgement rests on the record rather than on marketing.
Does Semgrep or SonarQube have more vulnerabilities?
SonarQube currently has 6 attributed advisories in this record and Semgrep has 0. That comparison is easy to misread. A vulnerability register records what researchers found and what the vendor disclosed, so a widely deployed product with an active disclosure process accumulates more entries than one nobody audits. A low count can equally mean a small install base or a vendor that publishes little. The two figures are also not measured with equal precision: an advisory only counts here when a register ties it to a named product, and 3 entries naming Semgrep and 16 naming SonarQube could not be tied to one. The count worth reading closely is how many entries appear in the CISA catalogue of known exploited vulnerabilities, which is 0 for Semgrep and 0 for SonarQube.
How independent is the evidence on Semgrep and SonarQube?
100 percent of the Semgrep record and 100 percent of the SonarQube record come from sources the vendor does not control: government vulnerability registers, public practitioner threads and a fixed list of recognised analysts and technology press. Material either vendor publishes about itself stays in the record and is labelled as a claim.
What does this comparison of Semgrep and SonarQube leave out?
Pricing, contract terms, notice periods, support commitments and the current scope of any certificate. None of those appear in a public register; they come from the vendor and belong in a negotiation record. Fit with your existing stack and your compliance scope is also absent, because it is a property of your organisation rather than of a devsecops and application security product.
Why does nexalign publish this instead of ranking Semgrep and SonarQube?
Because a ranking would have to invent the buyer. Every quadrant and score sheet silently assumes a weighting that belongs to someone else's decision. nexalign publishes the record and keeps the weighting where it belongs, in DecisionOS, where it is set by the people who have to defend the outcome.
The full records
Semgrep: the full evidence record
24 items from 3 kinds of source
SonarQube: the full evidence record
32 items from 5 kinds of source
Two records do not make a decision
Choosing between Semgrep and SonarQube depends on which criteria matter to you, how heavily each one weighs, what your hard constraints are and who has to sign the result off. DecisionOS takes this same evidence pool, scores it against the criteria of your actual decision and produces a memo that holds up in front of a board and an auditor.