Re: ANNOUNCE: shipcheck - CRA compliance auditor for Yocto builds

Javier Tia <[email protected]> Fri, 08 May 2026 19:30:03 -0600
Newsgroups org.yoctoproject.lists.yocto,org.openembedded.lists.openembedded-core
Message-ID <[email protected]>
Hi Olivier,

On 4/27/26 08:27 AM, Olivier Benjamin wrote:
> I would certainly be interested in being able to validate SPDX 3.0
> output.

That shipped in v0.0.6. The short version: Yocto Scarthgap scores
20/50 on the new check because create-spdx-3.0 doesn't emit supplier
or per-Package checksums in any form -- field or Relationship. The
format detection, CreationInfo, and rootElement checks all pass; it's
the 30 per-Package points that are zero.

Full details, the scoring breakdown, and a drafted 2-patch series for
openembedded-core to fix the emission gap are in the GitHub issue:

  https://github.com/jetm/shipcheck/issues/3

If you have a Scarthgap or walnascar build with create-spdx-3.0
enabled, I'd be curious what score you see:

  pip install shipcheck==0.0.6
  shipcheck check --build-dir <your-build-dir>

The sbom-generation row is the one to watch. Any feedback on the BSI
v2.1.0 -> SPDX 3.0 field mapping (committed at
audits/0003-spdx3-mapping/mapping.md, marked draft pending review)
would also be welcome -- you have more context on how the SPDX
community expects BSI's requirements to map to 3.0 constructs.

> Not super relevant, but I would dispute the "paperwork regulation"
> bit, and one can only gloss over the "scanner-selection" issue if
> one assumes that problem already solved.

Fair pushback. I was shortcutting. The paperwork framing was meant to
highlight the documentation gap, not to minimize the vulnerability-
management side of things.

-- 
Javier Tia