Repository navigation
Open
Labels
securityA credential, an exposure, or a control that is weaker than recordedA credential, an exposure, or a control that is weaker than recordedsensitivestacks/sensitive on trinity (VLAN 99)stacks/sensitive on trinity (VLAN 99)
Description
The weekly CVE scan found fixable HIGH or CRITICAL vulnerabilities in ghcr.io/immich-app/postgres. This body is rewritten by every scan, and the issue closes itself when a scan finds the pinned digest clean.
The fix is normally the Dependabot bump for this image. If upstream has not released one, the fixed versions below say what to wait for.
ghcr.io/immich-app/postgres:17-vectorchord0.4.3
Stacks: sensitive. Digest: sha256:79b95b795f0911e95ffa4d092a2162c4d81a984d5629bc8f786afe0bc0657671.
| Severity | ID | Package | Installed | Fixed in |
|---|---|---|---|---|
| CRITICAL | CVE-2026-33845 | libgnutls30 | 3.7.9-2+deb12u5 | 3.7.9-2+deb12u7 |
| CRITICAL | CVE-2026-42010 | libgnutls30 | 3.7.9-2+deb12u5 | 3.7.9-2+deb12u7 |
| CRITICAL | CVE-2026-13221 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-42496 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-8376 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-31789 | libssl3 | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| CRITICAL | CVE-2026-31789 | openssl | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| CRITICAL | CVE-2026-13221 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-42496 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-8376 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-13221 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-42496 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-8376 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-13221 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-42496 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| CRITICAL | CVE-2026-8376 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-109518 | apt | 2.6.1 | 2.6.2 |
| HIGH | CVE-2025-68973 | dirmngr | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gnupg | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gnupg-l10n | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gnupg-utils | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpg | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpg-agent | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpg-wks-client | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpg-wks-server | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpgconf | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpgsm | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2025-68973 | gpgv | 2.2.40-1.1 | 2.2.40-1.1+deb12u2 |
| HIGH | CVE-2026-109518 | libapt-pkg6.0 | 2.6.1 | 2.6.2 |
| HIGH | CVE-2026-4878 | libcap2 | 1:2.66-4+deb12u1 | 1:2.66-4+deb12u3 |
| HIGH | CVE-2026-33846 | libgnutls30 | 3.7.9-2+deb12u5 | 3.7.9-2+deb12u7 |
| HIGH | CVE-2026-3833 | libgnutls30 | 3.7.9-2+deb12u5 | 3.7.9-2+deb12u7 |
| HIGH | CVE-2026-42009 | libgnutls30 | 3.7.9-2+deb12u5 | 3.7.9-2+deb12u7 |
| HIGH | CVE-2026-40355 | libgssapi-krb5-2 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40356 | libgssapi-krb5-2 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40355 | libk5crypto3 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40356 | libk5crypto3 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40355 | libkrb5-3 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40356 | libkrb5-3 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40355 | libkrb5support0 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-40356 | libkrb5support0 | 1.20.1-2+deb12u3 | 1.20.1-2+deb12u5 |
| HIGH | CVE-2026-103111 | libpcre2-8-0 | 10.42-1 | 10.42-1+deb12u2 |
| HIGH | CVE-2026-86145 | libpcre2-8-0 | 10.42-1 | 10.42-1+deb12u1 |
| HIGH | CVE-2026-89157 | libpcre2-8-0 | 10.42-1 | 10.42-1+deb12u1 |
| HIGH | CVE-2026-89161 | libpcre2-8-0 | 10.42-1 | 10.42-1+deb12u1 |
| HIGH | CVE-2023-31484 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u3 |
| HIGH | CVE-2026-42497 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-48962 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57432 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57433 | libperl5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2025-6965 | libsqlite3-0 | 3.40.1-2+deb12u1 | 3.40.1-2+deb12u2 |
| HIGH | CVE-2025-15467 | libssl3 | 3.0.17-1~deb12u2 | 3.0.18-1~deb12u2 |
| HIGH | CVE-2025-69421 | libssl3 | 3.0.17-1~deb12u2 | 3.0.18-1~deb12u2 |
| HIGH | CVE-2026-28387 | libssl3 | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-28388 | libssl3 | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-28389 | libssl3 | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-28390 | libssl3 | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-45447 | libssl3 | 3.0.17-1~deb12u2 | 3.0.20-1~deb12u2 |
| HIGH | CVE-2026-84782 | libssl3 | 3.0.17-1~deb12u2 | 3.0.22-1~deb12u2 |
| HIGH | CVE-2025-7424 | libxslt1.1 | 1.1.35-1+deb12u1 | 1.1.35-1+deb12u2 |
| HIGH | CVE-2025-15467 | openssl | 3.0.17-1~deb12u2 | 3.0.18-1~deb12u2 |
| HIGH | CVE-2025-69421 | openssl | 3.0.17-1~deb12u2 | 3.0.18-1~deb12u2 |
| HIGH | CVE-2026-28387 | openssl | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-28388 | openssl | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-28389 | openssl | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-28390 | openssl | 3.0.17-1~deb12u2 | 3.0.19-1~deb12u2 |
| HIGH | CVE-2026-45447 | openssl | 3.0.17-1~deb12u2 | 3.0.20-1~deb12u2 |
| HIGH | CVE-2026-84782 | openssl | 3.0.17-1~deb12u2 | 3.0.22-1~deb12u2 |
| HIGH | CVE-2023-31484 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u3 |
| HIGH | CVE-2026-42497 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-48962 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57432 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57433 | perl | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2023-31484 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u3 |
| HIGH | CVE-2026-42497 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-48962 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57432 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57433 | perl-base | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2023-31484 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u3 |
| HIGH | CVE-2026-42497 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-48962 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57432 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
| HIGH | CVE-2026-57433 | perl-modules-5.36 | 5.36.0-7+deb12u2 | 5.36.0-7+deb12u4 |
Written by .github/workflows/cve-scan.yml in this run.
Metadata
Metadata
Assignees
Labels
securityA credential, an exposure, or a control that is weaker than recordedA credential, an exposure, or a control that is weaker than recordedsensitivestacks/sensitive on trinity (VLAN 99)stacks/sensitive on trinity (VLAN 99)
Activity
Gerrrt commentedon Oct 7, 2026
Triage, 2026-10-07: wait for Immich to move past VectorChord 0.4.3
Why there's no plain bump.
ghcr.io/immich-app/postgres:17-vectorchord0.4.3was built on 2025-10-07, and upstream has not rebuilt it since. The only way to newer packages is a newer VectorChord tag. Scanned with trivy 0.75.0 usingcve-scan.yml's flags:17-vectorchord0.4.3(pinned)17-vectorchord0.5.317-vectorchord1.1.1Immich would accept the newer tags. Immich 3.2.4's
VECTORCHORD_VERSION_RANGEis>=0.3 <2. On start it runsALTER EXTENSION vchord UPDATEand reindexes the CLIP and face vectors itself.Decided: wait. Immich's own
docker-compose.yml, at v3.2.4, at v3.3.0-rc.2 and onmain, still ships VectorChord 0.4.3. Moving the database that holds the photo library ahead of what upstream runs, for a partial reduction, isn't worth it. The database is reachable only from the Immich server inside the sensitive tier.Reopen this decision when:
🤖 Triage by Claude Code
Gerrrt commentedon Oct 7, 2026
Still waiting, re-checked 2026-10-07 evening
The trigger has not fired. Immich v3.3.0 shipped at 18:37 UTC today. Its
docker/docker-compose.ymlstill pinsghcr.io/immich-app/postgres:14-vectorchord0.4.3-pgvectors0.2.0, as v3.2.4 andmaindo. So the decision above stands.#1061 moved the pin to
17-vectorchord1.1.1ahead of upstream. It was opened without this triage in view, and it is closed unmerged in favour of it.Ready for when the trigger fires. #1061 ran the upgrade path on throwaway containers under
immich-db's exact hardening:cap_drop: ALL,no-new-privileges;--network none.The steps:
vchordrqcosine index and 500 rows.17-vectorchord1.1.1@sha256:7f0b786ed5a7a9d26522dacbf46114a76e043a515ce5e020690fc6702e987b72, with the same volume.ALTER EXTENSION vchord UPDATE TO '1.1.1', which is Immich's own startup step.REINDEX, after which an index-forced nearest-neighbour query returned 5 rows. The logs were clean.When Immich's compose moves past 0.4.3, the move is the pin plus the comment from #1061's diff. Take a
make backupset withimmich-dbfirst, and allow time for the reindex of the real CLIP and face indexes.🤖 Triage by Claude Code
[-]Fixable CVEs in ghcr.io/immich-app/postgres (20 critical, 119 high)[/-][+]Fixable CVEs in ghcr.io/immich-app/postgres (20 critical, 122 high)[/+]github-actions commentedon Oct 10, 2026
New since the last scan: CVE-2026-78667, CVE-2026-78669, CVE-2026-97031.
Gerrrt commentedon Oct 10, 2026
Still waiting, re-checked 2026-10-10
The trigger has not fired. Immich v3.3.1 shipped on 2026-10-08. Its
docker/docker-compose.ymlstill pinsghcr.io/immich-app/postgres:14-vectorchord0.4.3-pgvectors0.2.0, and so doesmain.There is no digest bump either. Upstream has not rebuilt
17-vectorchord0.4.3. The tag still resolves tosha256:79b95b795f0911e95ffa4d092a2162c4d81a984d5629bc8f786afe0bc0657671, which is the digest we pin.Nothing has changed in reachability.
immich-dbis still reachable only from the Immich server inside the sensitive tier. The decision and the triggers for revisiting it stand as written above, and the upgrade path #1061 tested is ready when they fire.🤖 Triage by Claude Code
chore(sensitive): Immich v3.2.4 -> v3.3.1, ahead of Sunday's Dependab…
Gerrrt commentedon Oct 10, 2026
This is waiting on upstream. Nothing in the repo needs to change.
ghcr.io/immich-app/postgres:17-vectorchord0.4.3(sha256:79b95b79…).pin-digests.shreports no drift, so the tag hasn't been rebuilt since it was pinned.stdlib: 59 findings, fixed in 27 different versions (see the table)libperl5.36: 8 findings, fixed in 2 different versions (see the table)libssl3: 8 findings, fixed in 3 different versions (see the table)openssl: 8 findings, fixed in 3 different versions (see the table)perl: 8 findings, fixed in 2 different versions (see the table)perl-base: 8 findings, fixed in 2 different versions (see the table)When upstream rebuilds or ships a release with these fixes, Dependabot should open the bump. Check it with trivy and
check_image_pins.py, then deploy. The next scan will then close this issue.Gerrrt commentedon Oct 10, 2026
The criticals, traced against the live database, 2026-10-10
This is the same trace as on #978, #979, #1002 and the others. I checked it against the running
sensitive-immich-dbon trinity (postgres:17-vectorchord0.4.3@sha256:79b95b79…); a fresh trivy 0.75.0 scan agrees. The 20 critical rows are 10 CVEs, and none is reachable. The Postgres settings below were read withpsqlinside the container (SHOW ssl,pg_hba_file_rules,pg_extension,pg_language).libldap-2.5(ldd:libldap-2.5.so.0 -> gnutls). Postgres does its own TLS with OpenSSL. libldap would use GnuTLS only as an LDAP client, and only forldapauth.pg_hbahas noldaprules:truston the local socket, 127.0.0.1 and ::1, andscram-sha-256for everything else. Nothing speaks DTLS.SHOW sslisoff, so Postgres negotiates no TLS at all, and the container isx86_64.docker-init → immich-docker-entrypoint.sh, then the postmaster and its background workers. No PL/Perl:pg_languageisc, internal, plpgsql, sql, and noplperl*.sois installed.Archive::Tarextracting a hostile tarballArchive/Tar.pmis present here, unlike in linkding and ML, but nothing invokes Perl./usr/local/bin/gosunet/netip, or crypto/tls session resumptiongosuis the entrypoint's privilege drop, used only when it starts as root. The container runs as uid 999 (postgres), so it's never called. Even when run, it does none of those things: itsetuids andexecs.Extensions in use are
vchord 0.4.3,vector 0.8.1,cube,earthdistance,pg_trgm,unaccent,uuid-osspandplpgsql, withshared_preload_libraries = vchord.so. None is implicated by any of the ten.This confirms the wait decided above, now with a trace behind it. The DB's network clients must pass
scram-sha-256, and the only one is immich-server inside the sensitive tier. The HIGHs aren't covered by this trace. The upgrade path #1061 tested (VectorChord 1.1.1, 13C/41H) stays parked until Immich's own compose moves.What would change this:
ssl, LDAP auth inpg_hba, orplperl.To reproduce:
🤖 Traced by Claude Code
feat(cve-scan): the gosu skip is a rule in gosu-skip.txt, and Immich'…
[-]Fixable CVEs in ghcr.io/immich-app/postgres (20 critical, 122 high)[/-][+]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 67 high)[/+][-]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 67 high)[/-][+]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 69 high)[/+]github-actions commentedon Oct 26, 2026
New since the last scan: CVE-2026-84782.
[-]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 69 high)[/-][+]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 71 high)[/+]github-actions commentedon Nov 9, 2026
New since the last scan: CVE-2026-109518.
Gerrrt commentedon Nov 9, 2026
Still waiting, re-checked 2026-11-09
Trigger 1 has not fired. Immich's
docker/docker-compose.ymlstill pinsghcr.io/immich-app/postgres:14-vectorchord0.4.3-pgvectors0.2.0at v3.3.6 (Latest, the version we run), at v3.4.0-rc.3 and onmain.VECTORCHORD_VERSION_RANGEis still>=0.3 <2.There is no rebuild either.
17-vectorchord0.4.3still resolves tosha256:79b95b795f0911e95ffa4d092a2162c4d81a984d5629bc8f786afe0bc0657671, which is the digest we pin. So Dependabot has no bump to open.Trigger 2 has not fired. The scan has added two CVEs since the 2026-10-10 trace, both HIGH. I checked them against the running
sensitive-immich-db, and neither is reachable:SHOW sslis stilloff, so no TLS is negotiated at all.ReadonlyRootfs=true), and its processes aredocker-init → immich-docker-entrypoint.sh, the postmaster and its background workers.The rest is unchanged.
pg_hbaauth methods are stilltrust(local) andscram-sha-256, and the image isx86_64.The decision and its triggers stand. The #1061 upgrade path (VectorChord 1.1.1) stays parked until Immich's own compose moves. When it does, take a
make backupset withimmich-dbfirst and allow time for the reindex.🤖 Triage by Claude Code
[-]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 71 high)[/-][+]Fixable CVEs in ghcr.io/immich-app/postgres (16 critical, 67 high)[/+]