ENH: Add vulnerability scan workflow with grype by matthewfeickert · Pull Request #164 · scientific-python/upload-nightly-action

matthewfeickert · 2026-04-23T07:15:45Z

Add GitHub Actions workflow that runs a vulnerability scan with grype (https://github.com/anchore/grype) on the conda packages in the Pixi environment.

HT to @wolfv and @pavelzw for the 2026-04-21 prefix.dev blog post Securing the Conda-Forge Supply Chain.

matthewfeickert · 2026-04-23T07:18:08Z

+          locked: true
+
+      - name: Scan with grype
+        run: pixi exec grype .pixi/envs/default


There is the flag

-f, --fail-on string set the return code to 2 if a vulnerability is found with a severity >= the given severity, options=[negligible low medium high critical]

but I decided to not have things fail at the moment as we have some critical ones that I don't think we can do much about yet.

$ pixi exec grype .pixi/envs/default --fail-on high | grep 'High\|Critical' ✔ Indexed file system .pixi/envs/default ✔ Cataloged contents 3913d5d0159f235532a38c6542273f76b141335ceeef368779ba833d9fbcb4ac ├── ✔ Packages [211 packages] ├── ✔ Executables [262 executables] ├── ✔ File metadata [11,099 locations] └── ✔ File digests [11,099 files] ✘ Scan for vulnerabilities [58 vulnerability matches] ├── by severity: 2 critical, 22 high, 27 medium, 4 low, 1 negligible (2 unknown) └── by status: 50 fixed, 8 not-fixed, 0 ignored [0000] WARN no explicit name and version provided for directory source, deriving artifact ID from the given path (which is not ideal) from=syft A newer version of grype is available for download: 0.111.1 (installed version is 0.111.0) [0002] ERROR discovered vulnerabilities at or above the severity threshold libtiff 4.7.1 conda CVE-2023-52356 High 0.8% (74th) 0.6 openssl 3.6.0 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1 conda CVE-2025-15467 High 0.7% (72nd) 0.6 openssl 3.6.0 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ... conda CVE-2025-69420 High 0.3% (53rd) 0.2 python 3.14.2 conda CVE-2026-6100 Critical 0.2% (35th) 0.1 jaraco-context 5.3.0 6.1.0 python GHSA-58pv-8j8x-9vj2 High < 0.1% (25th) < 0.1 libpng 1.6.53 1.6.55 conda CVE-2026-25646 High < 0.1% (23rd) < 0.1 openssl 3.6.0 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ... conda CVE-2025-69419 High < 0.1% (19th) < 0.1 openssl 3.6.0 3.6.2 conda CVE-2026-28386 Critical < 0.1% (15th) < 0.1 openssl 3.6.0 1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ... conda CVE-2026-28389 High < 0.1% (15th) < 0.1 openssl 3.6.0 1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ... conda CVE-2026-28390 High < 0.1% (15th) < 0.1 libpng 1.6.53 1.6.56 conda CVE-2026-33636 High < 0.1% (15th) < 0.1 libpng 1.6.53 1.6.56 conda CVE-2026-33416 High < 0.1% (11th) < 0.1 openssl 3.6.0 1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ... conda CVE-2025-69421 High < 0.1% (10th) < 0.1 libtiff 4.7.1 conda CVE-2026-4775 High < 0.1% (9th) < 0.1 urllib3 2.6.2 2.6.3 python GHSA-38jv-5279-wg99 High < 0.1% (8th) < 0.1 libpng 1.6.53 1.6.54 conda CVE-2026-22695 High < 0.1% (8th) < 0.1 openssl 3.6.0 1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ... conda CVE-2026-28388 High < 0.1% (7th) < 0.1 pillow 12.0.0 12.1.1 python GHSA-cfh3-3jmp-rvhc High < 0.1% (5th) < 0.1 pillow 12.0.0 12.2.0 python GHSA-whj4-6x5x-4v2j High < 0.1% (4th) < 0.1 python 3.14.2 conda CVE-2026-4786 High < 0.1% (5th) < 0.1 libpng 1.6.53 1.6.54 conda CVE-2026-22801 High < 0.1% (4th) < 0.1 openssl 3.6.0 3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2 conda CVE-2026-31790 High < 0.1% (4th) < 0.1 openssl 3.6.0 3.5.6, *3.6.2 conda CVE-2026-2673 High < 0.1% (4th) < 0.1 wheel 0.45.1 0.46.2 python GHSA-8rrh-rw8j-w5fx High < 0.1% (1st) < 0.1

Not sure what people think.

When there is a safe version I would fail. Or you're saying we can't do it here??

If we don't fail I am not sure I understand what this action does. People won't really have a look then no?

When there is a safe version I would fail.

@tupui so you're saying as there is the flag

--only-fixed ignore matches for vulnerabilities that are not fixed

to do

Suggested change

run: pixi exec grype .pixi/envs/default

run: pixi exec grype .pixi/envs/default --only-fixed --fail-on high

?

There's some information here that we don't want though / should filter out, as

$ pixi exec grype --only-fixed --fail-on high .pixi/envs/default | grep 'High\|Crit' ✔ Indexed file system .pixi/envs/default ✔ Cataloged contents 3913d5d0159f235532a38c6542273f76b141335ceeef368779ba833d9fbcb4ac ├── ✔ Packages [211 packages] ├── ✔ Executables [262 executables] ├── ✔ File metadata [11,099 locations] └── ✔ File digests [11,099 files] ✘ Scan for vulnerabilities [50 vulnerability matches] ├── by severity: 2 critical, 23 high, 27 medium, 4 low, 1 negligible (2 unknown) └── by status: 50 fixed, 9 not-fixed, 9 ignored [0000] WARN no explicit name and version provided for directory source, deriving artifact ID from the given path (which is not ideal) from=syft A newer version of grype is available for download: 0.111.1 (installed version is 0.111.0) [0002] ERROR discovered vulnerabilities at or above the severity threshold openssl 3.6.0 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1 conda CVE-2025-15467 High 0.7% (72nd) 0.6 openssl 3.6.0 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ... conda CVE-2025-69420 High 0.3% (53rd) 0.2 jaraco-context 5.3.0 6.1.0 python GHSA-58pv-8j8x-9vj2 High < 0.1% (25th) < 0.1 libpng 1.6.53 1.6.55 conda CVE-2026-25646 High < 0.1% (23rd) < 0.1 openssl 3.6.0 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ... conda CVE-2025-69419 High < 0.1% (19th) < 0.1 openssl 3.6.0 3.6.2 conda CVE-2026-28386 Critical < 0.1% (15th) < 0.1 openssl 3.6.0 1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ... conda CVE-2026-28389 High < 0.1% (15th) < 0.1 openssl 3.6.0 1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ... conda CVE-2026-28390 High < 0.1% (15th) < 0.1 libpng 1.6.53 1.6.56 conda CVE-2026-33636 High < 0.1% (15th) < 0.1 libpng 1.6.53 1.6.56 conda CVE-2026-33416 High < 0.1% (11th) < 0.1 openssl 3.6.0 1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ... conda CVE-2025-69421 High < 0.1% (10th) < 0.1 urllib3 2.6.2 2.6.3 python GHSA-38jv-5279-wg99 High < 0.1% (8th) < 0.1 libpng 1.6.53 1.6.54 conda CVE-2026-22695 High < 0.1% (8th) < 0.1 openssl 3.6.0 1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ... conda CVE-2026-28388 High < 0.1% (7th) < 0.1 pillow 12.0.0 12.1.1 python GHSA-cfh3-3jmp-rvhc High < 0.1% (5th) < 0.1 pillow 12.0.0 12.2.0 python GHSA-whj4-6x5x-4v2j High < 0.1% (4th) < 0.1 libpng 1.6.53 1.6.54 conda CVE-2026-22801 High < 0.1% (4th) < 0.1 openssl 3.6.0 3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2 conda CVE-2026-31790 High < 0.1% (4th) < 0.1 openssl 3.6.0 3.5.6, *3.6.2 conda CVE-2026-2673 High < 0.1% (4th) < 0.1 wheel 0.45.1 0.46.2 python GHSA-8rrh-rw8j-w5fx High < 0.1% (1st) < 0.1

has things like

NAME INSTALLED FIXED IN TYPE VULNERABILITY SEVERITY EPSS RISK ... jaraco-context 5.3.0 6.1.0 python GHSA-58pv-8j8x-9vj2 High < 0.1% (25th) < 0.1

which is a CVE for Python and not for the conda-forge package. So we want to also be able to ask it only for conda-forge related vulnerabilities that have a release with a fix.

--only-fixed --fail-on high

Ah yes I would try to do that, or at least on criticals?

--only-fixed

More of an open remark. In general if we get notified that something hit the fan really bad, we could consider having a way to stop the distribution completely.

@tupui I now have two checks, where the second is checking for the fixed and unfixed but fails only for critical vulnerabilities.

- name: Scan with grype (for fixed vulnerabilities) run: pixi exec grype .pixi/envs/default --only-fixed --fail-on high - name: Scan with grype (for all vulnerabilities) run: pixi exec grype .pixi/envs/default --fail-on critical

Ah, no, I had to flip them to

- name: Scan with grype (for all vulnerabilities) run: pixi exec grype .pixi/envs/default --fail-on critical - name: Scan with grype (for fixed vulnerabilities) run: pixi exec grype .pixi/envs/default --only-fixed --fail-on high

or else we would never pick up the unfixed critical vulnerabilities that you were mentioning checking for if there were known fixed vulnerabilities that were at high or critical.

So if we can pass on the most restrictive cut level (Critical) for everything, then we loosen the cut level to High and then check again for ones with fixes.

I guess we can make some arguments for both 😅

@tupui I think if we want to be able to test both and know if there are unfixed VeryBad ™️ vulnerabilities that still persist after the 30 dependency cooldown

upload-nightly-action/pixi.toml

Lines 9 to 10 in 7c719ad

# month window to find security issues

exclude-newer = "30d"

then we need to use what is currently there:

--fail-on critical

--only-fixed --fail-on high

matthewfeickert · 2026-04-23T07:18:29Z

@scientific-python/nightly-wheels-developers this is ready for review.

tupui

Nice! Did not know this one and I see they even support SBOM directly.

pavelzw · 2026-04-23T09:37:21Z

+  - package:
+      # only show vulnerabilities from conda-forge conda packages
+      type: python


i don't think this is what you want.
grype doesn't map type conda vulnerabilities (like

jaraco.context 6.1.1 pyhcf101f3_0 15.01 KiB conda https://conda.anaconda.org/conda-forge

)

to CVEs (like

jaraco-context 5.3.0 6.1.0 python GHSA-58pv-8j8x-9vj2 High < 0.1% (25th) < 0.1

). right now, there is no mapping for this. grype finds the typical site packages structure and infers the corresponding pypi package through this. if you excluded type: python packages, you would miss all python CVEs

see also https://pixi.prefix.dev/latest/security/#current-gaps-and-practical-recommendation

We are currently working on a PURL-related Conda Enhancement Proposal, conda/ceps#63, that will make it easier to match conda-installed software against CVEs that are tracked in other ecosystems like PyPI. Until that work is standardized and widely implemented, direct scans of the already-installed environment with tools like Syft and Grype remain the most practical workaround.

Thanks for explaining this in detail, @pavelzw.

matthewfeickert · 2026-04-23T15:22:42Z

Note that there's a lot of vulnerabilities coming from openssl

NAME            INSTALLED  FIXED IN                                             TYPE    VULNERABILITY        SEVERITY    EPSS           RISK   
openssl         3.6.0      3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1                  conda   CVE-2025-15467       High        0.7% (72nd)    0.6    
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    conda   CVE-2025-69420       High        0.3% (53rd)    0.2

but these won't be fixed anytime soon as conda-forge's global pinning are currently pinned to

openssl:
  - '3.5'

given a desire to stay on the LTS release in conda-forge/conda-forge-pinning-feedstock#7862.

edit: That being said, not sure how were getting openssl of v3.6.0 then at all, given that the 3.5 pin has been there for 5 months and I assume Python has been built again since then.

stefanv · 2026-04-23T21:25:04Z

Is the overarching goal of this PR to prevent uploading nightlies built with known-vulnerable build dependencies?

matthewfeickert · 2026-04-23T22:49:44Z

Is the overarching goal of this PR to prevent uploading nightlies built with known-vulnerable build dependencies?

@stefanv No. This doesn't give us any information on the uploaded packages or their build process — we only get them once they're already wheels. This only checks to see if the tool doing the uploading, anaconda-client, and its dependencies have any known vulnerabilities. That's the only thing that we can say doing security scans on software environments provides us, but, a compromised environment could effectively result in undefined behavior. We could imagine that in a compromised environment the upload process is mutating the packages being uploaded in a malicious manner (we don't have anything like Trusted Publishers or attestations for uploading to anaconda.org). Monitoring the known security vulnerabilities that exist in the environment allows for the possibility of risk assessment of using our provided software environment (by us or others).

matthewfeickert · 2026-04-27T07:23:03Z

@tupui (I'm no longer in European timezones, so I'll be slower to reply this week, but) do you have any other thoughts on this PR?

tupui

LGTM 🙌 (I am not the fastest as well these days, no worries 😉)

The one thing we could add somewhere is instructions on what to do if failing. Not everyone will know I suppose.

matthewfeickert · 2026-04-27T19:01:08Z

The one thing we could add somewhere is instructions on what to do if failing. Not everyone will know I suppose.

Good point. Let me add some docs.

stefanv · 2026-04-27T19:47:17Z

Failing because:

Anaconda token missing in CI/test upload.
Threshold exceeded for vulnerability check:

Details

[0032] ERROR discovered vulnerabilities at or above the severity threshold
NAME            INSTALLED  FIXED IN                                             TYPE    VULNERABILITY        SEVERITY    EPSS           RISK   
libtiff         4.7.1                                                           conda   CVE-2023-6277        Medium      3.8% (88th)    2.2    
openssl         3.6.0      3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1                  binary  CVE-2025-15467       High        0.7% (72nd)    0.6    
openssl         3.6.0      3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1                  conda   CVE-2025-15467       High        0.7% (72nd)    0.6    
libtiff         4.7.1                                                           conda   CVE-2023-52356       High        0.7% (72nd)    0.6    
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    binary  CVE-2025-69420       High        0.3% (53rd)    0.2    
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    conda   CVE-2025-69420       High        0.3% (53rd)    0.2    
python          3.14.2                                                          binary  CVE-2026-6100        Critical    0.2% (35th)    0.1    
python          3.14.2                                                          conda   CVE-2026-6100        Critical    0.2% (35th)    0.1    
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     binary  CVE-2026-0672        Medium      0.2% (36th)    < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     conda   CVE-2026-0672        Medium      0.2% (36th)    < 0.1  
jaraco-context  5.3.0      6.1.0                                                python  GHSA-58pv-8j8x-9vj2  High        < 0.1% (25th)  < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     binary  CVE-2026-0865        Medium      0.1% (32nd)    < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     conda   CVE-2026-0865        Medium      0.1% (32nd)    < 0.1  
libpng          1.6.53     1.6.55                                               conda   CVE-2026-25646       High        < 0.1% (23rd)  < 0.1  
openssl         3.6.0      1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ...  binary  CVE-2026-22796       Medium      0.1% (30th)    < 0.1  
openssl         3.6.0      1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ...  conda   CVE-2026-22796       Medium      0.1% (30th)    < 0.1  
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    binary  CVE-2025-69419       High        < 0.1% (19th)  < 0.1  
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    conda   CVE-2025-69419       High        < 0.1% (19th)  < 0.1  
python          3.14.2     3.15.0a6                                             binary  CVE-2025-15366       Medium      < 0.1% (23rd)  < 0.1  
python          3.14.2     3.15.0a6                                             binary  CVE-2025-15367       Medium      < 0.1% (23rd)  < 0.1  
python          3.14.2     3.15.0a6                                             conda   CVE-2025-15366       Medium      < 0.1% (23rd)  < 0.1  
python          3.14.2     3.15.0a6                                             conda   CVE-2025-15367       Medium      < 0.1% (23rd)  < 0.1  
openssl         3.6.0      3.6.2                                                binary  CVE-2026-28386       High        < 0.1% (16th)  < 0.1  
openssl         3.6.0      3.6.2                                                conda   CVE-2026-28386       High        < 0.1% (16th)  < 0.1  
python          3.14.2                                                          binary  CVE-2026-3298        High        < 0.1% (15th)  < 0.1  
python          3.14.2                                                          conda   CVE-2026-3298        High        < 0.1% (15th)  < 0.1  
libpng          1.6.53     1.6.56                                               conda   CVE-2026-33636       High        < 0.1% (15th)  < 0.1  
openssl         3.6.0      3.3.6, 3.4.4, 3.5.5, *3.6.1                          binary  CVE-2025-66199       Medium      < 0.1% (20th)  < 0.1  
openssl         3.6.0      3.3.6, 3.4.4, 3.5.5, *3.6.1                          conda   CVE-2025-66199       Medium      < 0.1% (20th)  < 0.1  
python          3.14.2                                                          binary  CVE-2026-1502        Medium      < 0.1% (18th)  < 0.1  
python          3.14.2                                                          conda   CVE-2026-1502        Medium      < 0.1% (18th)  < 0.1  
libtiff         4.7.1                                                           conda   CVE-2026-4775        High        < 0.1% (11th)  < 0.1  
libpng          1.6.53     1.6.56                                               conda   CVE-2026-33416       High        < 0.1% (10th)  < 0.1  
openssl         3.6.0      1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ...  binary  CVE-2025-69421       High        < 0.1% (10th)  < 0.1  
openssl         3.6.0      1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ...  conda   CVE-2025-69421       High        < 0.1% (10th)  < 0.1  
openssl         3.6.0      1.1.1zg, 3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2, ...    binary  CVE-2026-28387       High        < 0.1% (9th)   < 0.1  
openssl         3.6.0      1.1.1zg, 3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2, ...    conda   CVE-2026-28387       High        < 0.1% (9th)   < 0.1  
openssl         3.6.0      3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2                  binary  CVE-2026-31789       Critical    < 0.1% (7th)   < 0.1  
openssl         3.6.0      3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2                  conda   CVE-2026-31789       Critical    < 0.1% (7th)   < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     binary  CVE-2025-15282       Medium      < 0.1% (13th)  < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     binary  CVE-2026-1299        Medium      < 0.1% (13th)  < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     conda   CVE-2025-15282       Medium      < 0.1% (13th)  < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     conda   CVE-2026-1299        Medium      < 0.1% (13th)  < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           binary  CVE-2026-3644        Medium      < 0.1% (12th)  < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           conda   CVE-2026-3644        Medium      < 0.1% (12th)  < 0.1  
urllib3         2.6.2      2.6.3                                                python  GHSA-38jv-5279-wg99  High        < 0.1% (8th)   < 0.1  
openssl         3.6.0      1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ...  binary  CVE-2026-28389       High        < 0.1% (8th)   < 0.1  
openssl         3.6.0      1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ...  binary  CVE-2026-28390       High        < 0.1% (8th)   < 0.1  
openssl         3.6.0      1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ...  conda   CVE-2026-28389       High        < 0.1% (8th)   < 0.1  
openssl         3.6.0      1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ...  conda   CVE-2026-28390       High        < 0.1% (8th)   < 0.1  
libpng          1.6.53     1.6.54                                               conda   CVE-2026-22695       High        < 0.1% (8th)   < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     binary  CVE-2025-11468       Medium      < 0.1% (11th)  < 0.1  
python          3.14.2     3.10.20, 3.11.15, 3.12.13, 3.13.12, *3.14.3, ...     conda   CVE-2025-11468       Medium      < 0.1% (11th)  < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           binary  CVE-2026-4224        Medium      < 0.1% (10th)  < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           conda   CVE-2026-4224        Medium      < 0.1% (10th)  < 0.1  
openssl         3.6.0      1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ...  binary  CVE-2026-28388       High        < 0.1% (5th)   < 0.1  
openssl         3.6.0      1.0.2zp, 1.1.1zg, 3.0.20, 3.3.7, 3.4.5, *3.6.2, ...  conda   CVE-2026-28388       High        < 0.1% (5th)   < 0.1  
pillow          12.0.0     12.1.1                                               python  GHSA-cfh3-3jmp-rvhc  High        < 0.1% (5th)   < 0.1  
python          3.14.2                                                          binary  CVE-2026-4786        High        < 0.1% (5th)   < 0.1  
python          3.14.2                                                          conda   CVE-2026-4786        High        < 0.1% (5th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           binary  CVE-2026-3446        Medium      < 0.1% (7th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           conda   CVE-2026-3446        Medium      < 0.1% (7th)   < 0.1  
libpng          1.6.53     1.6.54                                               conda   CVE-2026-22801       High        < 0.1% (4th)   < 0.1  
openssl         3.6.0      1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ...  binary  CVE-2025-68160       Medium      < 0.1% (7th)   < 0.1  
openssl         3.6.0      1.0.2zn, 1.1.1ze, 3.0.19, 3.3.6, 3.4.4, *3.6.1, ...  conda   CVE-2025-68160       Medium      < 0.1% (7th)   < 0.1  
pillow          12.0.0     12.2.0                                               python  GHSA-whj4-6x5x-4v2j  High        < 0.1% (3rd)   < 0.1  
openssl         3.6.0      3.5.6, *3.6.2                                        binary  CVE-2026-2673        High        < 0.1% (4th)   < 0.1  
openssl         3.6.0      3.5.6, *3.6.2                                        conda   CVE-2026-2673        High        < 0.1% (4th)   < 0.1  
openssl         3.6.0      3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2                  binary  CVE-2026-31790       High        < 0.1% (3rd)   < 0.1  
openssl         3.6.0      3.0.20, 3.3.7, 3.4.5, 3.5.6, *3.6.2                  conda   CVE-2026-31790       High        < 0.1% (3rd)   < 0.1  
cryptography    46.0.5     46.0.7                                               python  GHSA-p423-j2cm-9vmq  Medium      < 0.1% (5th)   < 0.1  
openssl         3.6.0      3.3.6, 3.4.4, 3.5.5, *3.6.1                          binary  CVE-2025-15468       Medium      < 0.1% (6th)   < 0.1  
openssl         3.6.0      3.3.6, 3.4.4, 3.5.5, *3.6.1                          conda   CVE-2025-15468       Medium      < 0.1% (6th)   < 0.1  
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    binary  CVE-2026-22795       Medium      < 0.1% (5th)   < 0.1  
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    conda   CVE-2026-22795       Medium      < 0.1% (5th)   < 0.1  
python          3.14.2                                                          binary  CVE-2026-6019        Low         < 0.1% (11th)  < 0.1  
python          3.14.2                                                          conda   CVE-2026-6019        Low         < 0.1% (11th)  < 0.1  
python          3.14.2                                                          binary  CVE-2025-12781       Medium      < 0.1% (4th)   < 0.1  
python          3.14.2                                                          conda   CVE-2025-12781       Medium      < 0.1% (4th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a7                           binary  CVE-2026-2297        Medium      < 0.1% (4th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a7                           conda   CVE-2026-2297        Medium      < 0.1% (4th)   < 0.1  
wheel           0.45.1     0.46.2                                               python  GHSA-8rrh-rw8j-w5fx  High        < 0.1% (1st)   < 0.1  
python-dotenv   1.2.1      1.2.2                                                python  GHSA-mf9w-mj56-hr94  Medium      < 0.1% (3rd)   < 0.1  
libpng          1.6.53     1.6.57                                               conda   CVE-2026-34757       Medium      < 0.1% (3rd)   < 0.1  
libtiff         4.7.1                                                           conda   CVE-2023-6228        Medium      < 0.1% (3rd)   < 0.1  
openssl         3.6.0      3.4.4, 3.5.5, *3.6.1                                 binary  CVE-2025-11187       Medium      < 0.1% (0th)   < 0.1  
openssl         3.6.0      3.4.4, 3.5.5, *3.6.1                                 conda   CVE-2025-11187       Medium      < 0.1% (0th)   < 0.1  
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    binary  CVE-2025-69418       Medium      < 0.1% (0th)   < 0.1  
openssl         3.6.0      1.1.1ze, 3.0.19, 3.3.6, 3.4.4, 3.5.5, *3.6.1, ...    conda   CVE-2025-69418       Medium      < 0.1% (0th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           binary  CVE-2025-13462       Low         < 0.1% (2nd)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           conda   CVE-2025-13462       Low         < 0.1% (2nd)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           binary  CVE-2026-4519        Low         < 0.1% (0th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           conda   CVE-2026-4519        Low         < 0.1% (0th)   < 0.1  
openssl         3.6.0      3.5.5, *3.6.1                                        binary  CVE-2025-15469       Medium      < 0.1% (0th)   < 0.1  
openssl         3.6.0      3.5.5, *3.6.1                                        conda   CVE-2025-15469       Medium      < 0.1% (0th)   < 0.1  
cryptography    46.0.5     46.0.6                                               python  GHSA-m959-cc7f-wv43  Low         < 0.1% (0th)   < 0.1  
requests        2.32.5     2.33.0                                               python  GHSA-gc5v-m9x4-r6x2  Medium      < 0.1% (0th)   < 0.1  
pygments        2.19.2     2.20.0                                               python  GHSA-5239-wwwm-4pmq  Low         < 0.1% (0th)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           binary  CVE-2026-3479        Negligible  < 0.1% (3rd)   < 0.1  
python          3.14.2     3.13.13, *3.14.4, 3.15.0a8                           conda   CVE-2026-3479        Negligible  < 0.1% (3rd)   < 0.1

tupui

Nice docs thanks 🙏

matthewfeickert · 2026-04-27T21:58:06Z

@stefanv Can I also get a review from you (no rush on this though)?

bsipocz

Overall this looks good to me; but I think we should really aim to get this job in a green passing state otherwise it will be very difficult to spot any new issues.

Getting the job to pass can be done in a follow-up if you prefer, so I'm OK with merging this as is.

matthewfeickert · 2026-04-30T21:42:39Z

I think we should really aim to get this job in a green passing state otherwise it will be very difficult to spot any new issues.

Getting the job to pass can be done in a follow-up if you prefer, so I'm OK with merging this as is.

@bsipocz I fully agree, but I want to be honest that I think that this probably won't happen for a long time unless there's some discussion with conda-forge people.

To avoid having a broken weekly CI test for a long time I could open up some discussion RE: the global pinning and see how this would also affect the conda-forge feedstocks for python and cryptography. Thoughts?

tupui · 2026-04-30T21:54:22Z

this probably won't happen for a long time unless there's some discussion with conda-forge people.

That's a bit outrageous they would take some sweet time to fix things considering all that is happening now in terms of supply chain hacks 😅

matthewfeickert · 2026-04-30T22:00:56Z

That's a bit outrageous they would take some sweet time to fix things considering all that is happening now in terms of supply chain hacks 😅

I'm not saying that they would be unhelpful. I'm saying that given the build system #164 (comment) I would probably need to ask on the conda-forge Zulip to have a member of conda-forge/core explain to me if the global pinning to openssl v3.5.x in terms of build requirements has a knock on effect of keeping v3.6.1 from getting isntalled, or if that is just the version that python and cryptography got built against and we need to talk to their maintainer teams to get rebuilds.

edit: Actually, from pixi browse -m python I can see that python's run requirement on openssl isn't constrained here

and from pixi browse -m cryptography neither is cryptography's

So I'm not sure what is going on.

matthewfeickert · 2026-04-30T22:20:32Z

So I'm not sure what is going on.

What was going on is me forgetting that if I run

pixi upgrade

and the thing I'm upgrading

upload-nightly-action/pixi.toml

Line 15 in 7c719ad

anaconda-client = "==1.14.1"

is already at its most recent release it isn't going to resolve the environment all the way down. If I resolve the lock file though after loosening the dependency cooldown (like @henryiii suggested) then we get

$ pixi list openssl
Name     Version  Build           Size  Kind   Source
openssl  3.6.2    h35e630c_0  3.02 MiB  conda  https://conda.anaconda.org/conda-forge

and the only remaining critical vulnerability is in Python

$ pixi run grype
✨ Pixi task (grype): pixi exec grype .pixi/envs/default: (Scan environment for security vulnerabilities with grype)
 ✔ Indexed file system                                                                                                                                             .pixi/envs/default 
 ✔ Cataloged contents                                                                                                3913d5d0159f235532a38c6542273f76b141335ceeef368779ba833d9fbcb4ac 
   ├── ✔ Packages                        [217 packages]  
   ├── ✔ Executables                     [279 executables]  
   ├── ✔ File metadata                   [11,892 locations]  
   └── ✔ File digests                    [11,892 files]  
 ✔ Scanned for vulnerabilities     [22 vulnerability matches]  
   ├── by severity: 2 critical, 6 high, 12 medium, 2 low, 0 negligible
   └── by status:   4 fixed, 18 not-fixed, 0 ignored 
[0000]  WARN no explicit name and version provided for directory source, deriving artifact ID from the given path (which is not ideal) from=syft
NAME     INSTALLED  FIXED IN  TYPE    VULNERABILITY   SEVERITY  EPSS           RISK   
libtiff  4.7.1                conda   CVE-2023-6277   Medium    3.8% (88th)    2.2    
libtiff  4.7.1                conda   CVE-2023-52356  High      0.7% (72nd)    0.6    
python   3.14.4               binary  CVE-2026-6100   Critical  0.2% (35th)    0.1    
python   3.14.4               conda   CVE-2026-6100   Critical  0.2% (35th)    0.1    
python   3.14.4               binary  CVE-2026-3298   High      < 0.1% (20th)  < 0.1  
python   3.14.4               conda   CVE-2026-3298   High      < 0.1% (20th)  < 0.1  
python   3.14.4     3.15.0a6  binary  CVE-2025-15366  Medium    < 0.1% (23rd)  < 0.1  
python   3.14.4     3.15.0a6  binary  CVE-2025-15367  Medium    < 0.1% (23rd)  < 0.1  
python   3.14.4     3.15.0a6  conda   CVE-2025-15366  Medium    < 0.1% (23rd)  < 0.1  
python   3.14.4     3.15.0a6  conda   CVE-2025-15367  Medium    < 0.1% (23rd)  < 0.1  
python   3.14.4               binary  CVE-2026-1502   Medium    < 0.1% (18th)  < 0.1  
python   3.14.4               conda   CVE-2026-1502   Medium    < 0.1% (18th)  < 0.1  
libtiff  4.7.1                conda   CVE-2026-4775   High      < 0.1% (11th)  < 0.1  
python   3.14.4               binary  CVE-2026-3087   Medium    < 0.1% (14th)  < 0.1  
python   3.14.4               conda   CVE-2026-3087   Medium    < 0.1% (14th)  < 0.1  
python   3.14.4               binary  CVE-2026-4786   High      < 0.1% (5th)   < 0.1  
python   3.14.4               conda   CVE-2026-4786   High      < 0.1% (5th)   < 0.1  
python   3.14.4               binary  CVE-2026-6019   Low       < 0.1% (15th)  < 0.1  
python   3.14.4               conda   CVE-2026-6019   Low       < 0.1% (15th)  < 0.1  
python   3.14.4               binary  CVE-2025-12781  Medium    < 0.1% (4th)   < 0.1  
python   3.14.4               conda   CVE-2025-12781  Medium    < 0.1% (4th)   < 0.1  
libtiff  4.7.1                conda   CVE-2023-6228   Medium    < 0.1% (3rd)   < 0.1

and as CVE-2026-6100 doesn't have a fix, we can't do anything about that yet.

matthewfeickert · 2026-04-30T22:31:57Z

+If there are updates available try resolving the lock file fresh with
+
+```
+pixi reinstall
+```
+
+or by
+
+```
+rm pixi.lock && pixi lock
+```


Gave instructions to avoid my mistake here now.

Okay, now finalized. Sorry for the noise.

* Add GitHub Actions workflow that runs a vulnerability scan with grype (https://github.com/anchore/grype) on the conda packages in the Pixi environment. Check for vulnerabilities across all installed packages, both those with and without fixes, and then check for vulnerabilities that do have known fixes as a more restrictive failure setting.

* Add a Pixi task that executes grype but with 'pixi exec' to make it transient and to avoid adding grype to the default environment.

bsipocz · 2026-04-30T22:46:29Z

I am mostly thinking along the line of if it's failing all the time, then we diff it to known issues, like all the ones you highlight?

(I just say this as I have a couple of places where failures are expected and are always around, I find it really really difficult to even looking at the logs carefully as I know they are expected to fail. And naturally new stuff if always burried in those logs. Anyway, I'm much better with the ones where it's just an occasional failure).

matthewfeickert self-assigned this Apr 23, 2026

matthewfeickert commented Apr 23, 2026

View reviewed changes

tupui reviewed Apr 23, 2026

View reviewed changes

matthewfeickert force-pushed the feat/add-grype branch from 81e579d to 4c827c2 Compare April 23, 2026 09:31

pavelzw reviewed Apr 23, 2026

View reviewed changes

Comment thread .github/workflows/grype.yml Outdated

matthewfeickert force-pushed the feat/add-grype branch from 4c827c2 to 6d52911 Compare April 23, 2026 09:31

matthewfeickert requested review from pavelzw and tupui April 23, 2026 09:32

pavelzw reviewed Apr 23, 2026

View reviewed changes

matthewfeickert force-pushed the feat/add-grype branch from 6d52911 to 69a1fb3 Compare April 23, 2026 13:49

matthewfeickert force-pushed the feat/add-grype branch from 69a1fb3 to ee4ddf9 Compare April 23, 2026 16:06

tupui approved these changes Apr 27, 2026

View reviewed changes

matthewfeickert requested a review from tupui April 27, 2026 19:20

tupui approved these changes Apr 27, 2026

View reviewed changes

bsipocz approved these changes Apr 30, 2026

View reviewed changes

matthewfeickert force-pushed the feat/add-grype branch from efb664d to c6fa3da Compare April 30, 2026 22:17

matthewfeickert requested review from bsipocz and tupui April 30, 2026 22:26

matthewfeickert force-pushed the feat/add-grype branch from c6fa3da to 0184cca Compare April 30, 2026 22:31

matthewfeickert commented Apr 30, 2026

View reviewed changes

matthewfeickert force-pushed the feat/add-grype branch from 0184cca to 84bc621 Compare April 30, 2026 22:37

matthewfeickert commented Apr 30, 2026

View reviewed changes

Comment thread pixi.toml Outdated

matthewfeickert added 5 commits April 30, 2026 16:40

MNT: Loosen dependency cooldown to 14 days

9785651

MNT: Update lock file

2ce962a

ENH: Add grype Pixi task

094b6e1

* Add a Pixi task that executes grype but with 'pixi exec' to make it transient and to avoid adding grype to the default environment.

DOC: Add MAINTAINERS.md with notes on grype security scans

0961bc5

matthewfeickert force-pushed the feat/add-grype branch from 84bc621 to 0961bc5 Compare April 30, 2026 22:40

	run: pixi exec grype .pixi/envs/default
	run: pixi exec grype .pixi/envs/default --only-fixed --fail-on high

Conversation

matthewfeickert commented Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

matthewfeickert Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewfeickert Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewfeickert Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewfeickert Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewfeickert commented Apr 23, 2026

Uh oh!

tupui left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

pavelzw Apr 23, 2026 • edited by matthewfeickert Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewfeickert commented Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

stefanv commented Apr 23, 2026

Uh oh!

matthewfeickert commented Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

matthewfeickert commented Apr 27, 2026

Uh oh!

tupui left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewfeickert commented Apr 27, 2026

Uh oh!

stefanv commented Apr 27, 2026

Uh oh!

tupui left a comment

Choose a reason for hiding this comment

Uh oh!

matthewfeickert commented Apr 27, 2026

Uh oh!

bsipocz left a comment

Choose a reason for hiding this comment

Uh oh!

matthewfeickert commented Apr 30, 2026

Uh oh!

tupui commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

matthewfeickert commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

matthewfeickert commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

matthewfeickert commented Apr 23, 2026 •

edited

Loading

matthewfeickert Apr 23, 2026 •

edited

Loading

matthewfeickert Apr 23, 2026 •

edited

Loading

matthewfeickert Apr 23, 2026 •

edited

Loading

matthewfeickert Apr 23, 2026 •

edited

Loading

pavelzw Apr 23, 2026 •

edited by matthewfeickert

Loading

matthewfeickert commented Apr 23, 2026 •

edited

Loading

matthewfeickert commented Apr 23, 2026 •

edited

Loading

tupui left a comment •

edited

Loading

tupui commented Apr 30, 2026 •

edited

Loading

matthewfeickert commented Apr 30, 2026 •

edited

Loading

matthewfeickert commented Apr 30, 2026 •

edited

Loading