Software Bill of Materials became one of the latest buzzword. A lot of people and companies talking about it like a magical thing, if you use it then all of your security problems will be solved, just like what happened with Blockchain!!.

Though a hand full of projects (or companies building those projects) focused on the actual tooling part. Things we can use and see some useful output than blogposts/presentations with fancy graphics.

In this post we will try to see how can we use these tools today (2023/09/20).

SBOM currently comes in two major flavors, SPDX aka Software Package Data Index and CycloneDX. There are existing tooling to convert in between.

Syft

We will use syft from Anchore to generate our SBOM(s).

This tool can generate from various sources, starting from container images to Python projects, RPM/Debian dbs, Rust or Go projects.

Let us generate the SBOM for a Debian 12 VM.

$ syft /var/lib/dpkg -o spdx-json=server.spdx.json --source-name debian12 
 ✔ Indexed file system                                                                                         /var/lib/dpkg
 ✔ Cataloged packages              [395 packages]  

For for a Rust project:

$ syft /home/kdas/code/johnnycanencrypt/Cargo.lock -o spdx-json=jce.spdx.json
 ✔ Indexed file system                                                                      /home/kdas/code/johnnycanencrypt
 ✔ Cataloged packages              [203 packages]

We generated the SBOMs. Now this should solve the security issues, isn't?

I found the above in Matthew Martin's timeline.

Grype

This is where Grype comes handy, it is a vulnerability scanner for container images and filesystems and works with the SBOM(s) generated by syft.

$ grype jce.spdx.json 
 ✔ Vulnerability DB                [updated]  
 ✔ Scanned for vulnerabilities     [1 vulnerability matches]  
   ├── by severity: 0 critical, 0 high, 1 medium, 0 low, 0 negligible
   └── by status:   1 fixed, 0 not-fixed, 0 ignored 
NAME  INSTALLED  FIXED-IN  TYPE        VULNERABILITY        SEVERITY 
time  0.1.45     0.2.23    rust-crate  GHSA-wcg3-cvx6-7396  Medium

And:

grype server.spdx.json 
 ✔ Vulnerability DB                [no update available]  
 ✔ Scanned for vulnerabilities     [178 vulnerability matches]  
   ├── by severity: 6 critical, 136 high, 34 medium, 2 low, 0 negligible
   └── by status:   0 fixed, 178 not-fixed, 0 ignored 
NAME     INSTALLED     FIXED-IN  TYPE  VULNERABILITY     SEVERITY 
file     1:5.44-3                      CVE-2007-1536     High      
git      1:2.39.2-1.1                  CVE-2020-5260     High      
gnupg    2.2.40-1.1                    CVE-2022-3515     Critical  
gnupg    2.2.40-1.1                    CVE-2022-34903    Medium    
gnupg    2.2.40-1.1                    CVE-2022-3219     Low       
openssl  3.0.9-1                       CVE-2023-4807     High      
openssl  3.0.9-1                       CVE-2023-3817     Medium    
openssl  3.0.9-1                       CVE-2023-2975     Medium    
openssl  3.0.9-1                       CVE-2023-1255     Medium    
perl     5.36.0-7                      CVE-2023-31486    High      
perl     5.36.0-7                      CVE-2023-31484    High      
vim      2:9.0.1378-2                  CVE-2022-3520     Critical  
vim      2:9.0.1378-2                  CVE-2022-0318     Critical  
vim      2:9.0.1378-2                  CVE-2017-6350     Critical  
vim      2:9.0.1378-2                  CVE-2017-6349     Critical  
vim      2:9.0.1378-2                  CVE-2017-5953     Critical  
vim      2:9.0.1378-2                  CVE-2023-4781     High      
vim      2:9.0.1378-2                  CVE-2023-4752     High      

<snipped>

Now it is on your team members to decide how to react to information we gather from these tools. The tools themselves will not solve the problems at hand. You have to decide the update steps and if that is at all required or not.

Also please remember, there is and will be a lot of false positives (not in Grype output yet, but other tools in the SBOM ecosystem). The projects (I am talking about in general most of the tooling in this field) are trying hard to reduce these, but not possible always to remove every such edge case.