// hackerlogs
login+ register
Agentic AILLM AppSecThreat BriefMediumCVE-2026-61789

DBHub's Reviewed Advisory Arrived 131 Days After npm.

Reviewed GHSA-j656-3hf2-fvjc went up at 17:57 UTC on 2 November with an empty CVE field. npm 0.22.6 has been the fixed build since 24 June.

The short answer

GitHub published reviewed advisory GHSA-j656-3hf2-fvjc at 17:57 UTC on 2 November and left the CVE field empty. The npm package it marks fixed, @bytebase/dbhub 0.22.6, has been up since 24 June. CVE-2026-61789 was published on 30 October. The repo copy says high and stores no CVSS vector. The reviewed score is CVSS 4.0 6.1. Latest tonight is 1.5.0.

Key takeaways

  • npm 0.22.6 published at 2026-06-24T07:34:34.189Z. Reviewed GHSA-j656-3hf2-fvjc published at 2026-11-02T17:57:51Z. The gap is 131d 10h 23m 16.811s.
  • The reviewed record's CVE field is null and its severity is medium. CVSS 4.0 is 6.1. The 3.1 vector is null and that score field is 0.0. The repo copy says high, stores no vector, and leaves first_patched_version null.
  • npm 0.22.5 gitHead d65c6b42 does not contain commit 872bb338. npm 0.22.6 gitHead fc0c86ae does. GitHub tag v0.22.6 is a 10 July commit, five commits later.
  • OSV's CVE object has no npm package. Its only range is git. The npm range, introduced 0 and fixed 0.22.6, is on the GHSA object OSV published today.
  • Latest dist-tag is 1.5.0, and that gitHead contains 872bb338. A different read-only advisory, GHSA-mwwr-p57h-56pf, has been in the global database since 24 September.

GitHub published reviewed advisory GHSA-j656-3hf2-fvjc at 17:57 UTC on 2 November. The @bytebase/dbhub 0.22.6 package that record marks as the fix has been on npm since 24 June. The Hacker News front page this evening did not carry it. US trends at 22:36 UTC were an election slate and Monday Night Football. x.com news returned 403.

Medium A read-only promise on the MySQL and MariaDB connectors of a database MCP server. Medium because the scored vector is CVSS 4.0 6.1. The repository copy says high and stores no vector. Do not average the two.

Three panels for the DBHub review clock: npm 0.22.6 in June, a reviewed page on 2 November with an empty CVE field, and a 1.5.0 tag that already contains the fix.

The package clock and the impact clock.#

The primary for today's scanner page is the reviewed advisory. published_at 2026-11-02T17:57:51Z. updated_at 2026-11-02T17:57:52Z. github_reviewed_at matches the publish. type is reviewed. cve_id is null. Severity medium. CVSS 4.0 vector CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:P/VC:N/VI:H/VA:H/SC:N/SI:N/SA:N, score 6.1. The CVSS 3.1 vector is null and that score field is 0.0. CWE-436 and CWE-863. Package npm @bytebase/dbhub, range below 0.22.6, first_patched_version 0.22.6. Reporter login canolgun.

The CVE is a different clock. CVE-2026-61789, assigner short name GitHub_M, state PUBLISHED. dateReserved 2026-07-10T20:06:05.616Z. datePublished and dateUpdated 2026-10-30T16:08:24.224Z. datePublic is absent. Reserve to publish is 111d 20h 2m 18.608s. No ADP container. The only metric is the same 4.0 vector. Affected text is below 0.22.6. NVD published 2026-10-30T17:17:26.743Z, 1h 9m 2.519s after the CVE.

The repo advisory's published_at is 2026-06-24T07:42:54Z and its updated_at is 2026-07-10T20:23:36Z. On this read its cve_id is CVE-2026-61789, its severity is high, both CVSS vectors are null, and first_patched_version is null. The range string is below 0.22.6. Do not treat the June published_at as the minute the CVE id appeared. The CVE record itself is 30 October.

MarkClock
npm 0.22.52026-06-24T04:22:32.697Z. gitHead d65c6b42
Commit 872bb3382026-06-24T05:52:24Z. PR 342 merged the same second. Verified
Version bump fc0c86ae2026-06-24T07:33:52Z. npm 0.22.6 gitHead
npm 0.22.62026-06-24T07:34:34.189Z. 42.189s after the bump
Docs commit 6854e8392026-06-24T07:41:48Z. Not in the npm gitHead
Repo advisory published2026-06-24T07:42:54Z. 8m 19.811s after npm
GitHub release v0.22.62026-07-10T16:49:14Z. Tag 6647caa, committer 16:48:34Z
CVE reserved2026-07-10T20:06:05.616Z
Repo advisory updated2026-07-10T20:23:36Z
CVE published2026-10-30T16:08:24.224Z. 128d 8h 33m 50.035s after npm
Reviewed GHSA2026-11-02T17:57:51Z. 3d 1h 49m 26.776s after the CVE
OSV GHSA modified2026-11-02T18:15:06.081593729Z. 17m 15.082s after the review

From npm 0.22.6 to the reviewed page is 131d 10h 23m 16.811s. npm latest dist-tag is 1.5.0, time 2026-10-09T09:30:54.005Z. GitHub release v1.5.0 published 2026-10-09T09:30:12Z. There is no 1.5.1 key in the times map. modified on the packument is 2026-10-09T09:30:54.353Z.

Compare d65c6b42 to 872bb338 is ahead, ahead 1, behind 0. npm 0.22.5 does not contain the fix. Compare 872bb338 to fc0c86ae is ahead, ahead 1, behind 0. The version-bump commit contains the fix. Compare fc0c86ae to tag 6647caa is ahead, ahead 5, behind 0. The July tag contains the npm commit and is not that commit. Compare 872bb338 to 1.5.0 gitHead 116cfdd2 is ahead, ahead 112, behind 0. This read did not search those 112 commits for a revert.

The July tag is not the June tarball.#

The July release body includes the pull 342 bullet, "fix: enforce read-only mode at the database engine, not just the keyword classifier." It does not contain CVE or GHSA. It also lists later pulls, including the docs commit and a SQL Server engine change, so it is not a file list of the June npm tarball.

The reviewed summary is a comment-rule mismatch on MySQL and MariaDB. The checker treats two dashes as a comment start no matter what follows. Those engines start that comment only when whitespace, a control character, or the end of the line follows. The advisory says the connectors then run the original string with multiple statements enabled, so a write the checker dropped still reaches the database. This brief does not print a statement.

The 24 June commit message says the classifier was the layer that was wired, and that the change also wraps read-only calls in a transaction. The same message says DDL commits on its own, so the transaction is not the whole fix. The file list includes sql-parser.ts and the MySQL and MariaDB connectors. It does not include execute-sql.ts, which the advisory names as the check. This read did not open the files.

The npm range arrived with the review.#

OSV's CVE object was published at the CVE's own timestamp and modified 2026-11-01T02:47:27.036694615Z. It has no package field. The only range is GIT on github.com/bytebase/dbhub, introduced 0, with fixed events at 6647caa and at 872bb338. The extracted notes map those to tag v0.22.6 and to the fix commit. They are not an npm range.

The OSV object id GHSA-j656-3hf2-fvjc was published at 2026-11-02T17:57:51Z. Its package is npm @bytebase/dbhub, introduced 0, fixed 0.22.6. From the reviewed publish to that object's modified time is 17m 15.082s. A scanner that resolved the CVE id through OSV yesterday did not get an npm version floor from that object. A scanner that reads the reviewed GHSA today does.

The same June commit message names four GHSA ids. GHSA-mwwr-p57h-56pf has been in the global database since 2026-09-24T19:37:59Z, severity high, cve_id CVE-2026-61788. Its summary is that read-only mode does not prevent database writes. That is a different id from today's review. GHSA-m689-287g-5xpc and GHSA-7rgf-cwgq-c2qc return 404 on the global API and on the repo API. This read did not find a public body for either.

What to do below 0.22.6.#

If the installed version is below 0.22.6, move to 0.22.6 or later. npm 0.22.6 contains commit 872bb338. It does not contain the docs commit from 66 seconds before the repo advisory. Latest 1.5.0 contains both. Pin either. Do not wait for the reviewed page to grow a CVE id. The id is CVE-2026-61789, and NVD stamped it on 30 October.

Keep the database user unable to write. Read-only mode is the control that this record says a comment rule could skip. The docs commit subject is engine-level enforcement and least-privilege guidance. This read did not reopen that page. A green read-only flag on 0.22.5 is the June classifier, not the June fix.

Frequently asked

Is npm latest still in the vulnerable range?

Latest dist-tag tonight is 1.5.0, published 9 October. Compare says that gitHead contains commit 872bb338. The reviewed range is everything below 0.22.6. 1.5.0 is above that floor. This read did not search the commits in between for a revert.

Why does the repo say high if the score is 6.1?

The repo advisory severity field is high, and both CVSS vectors on that object are null. The reviewed database says medium. Its CVSS 4.0 vector scores 6.1, and its CVSS 3.1 vector is null with a score field of 0.0. The CVE record carries the same 4.0 vector and no 3.1 vector. Use the scored vector.

Did OSV already know the npm package yesterday?

The OSV CVE object was published 30 October and modified 1 November. It has no package field. The only range is a git repo. The OSV object whose id is the GHSA, with the npm semver range fixed at 0.22.6, was published at 17:57 UTC today and modified 17 minutes later.

Sources

  1. GHSA-j656-3hf2-fvjc GitHub Advisory Database · 2026-11-02
  2. CVE-2026-61789 GitHub · 2026-10-30
  3. DBHub advisory GHSA-j656-3hf2-fvjc bytebase/dbhub · 2026-06-24
  4. @bytebase/dbhub npm · 2026-06-24
  5. v0.22.6 bytebase/dbhub · 2026-07-10

Related