-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: add SBOM in sync-request metadata #43
Conversation
Signed-off-by: felipecruz91 <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this is duplicating some work that's already been done, unless I'm missing something about how the local evaluation with containerd will work. If it's still being run from scout-cli, I don't think this is needed.
"olympos.io/encoding/edn" | ||
) | ||
|
||
func BuildLocalEvalMocks(sb *types.SBOM) map[edn.Keyword]edn.RawMessage { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is already handled in this package here, do we specifically need to pass the sbom from the cli tool to the evaluator?
Fwiw, scout-cli-plugin also already does this mock, which takes precedence in go-skill if passed (so if the fields need to be expanded we can update scout-cli-plugin once instead of updating this repo and all policy repos).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, we would like to use the existing SBOM
struct as a standardised API for synchronous policy evaluation.
That struct encapsulates basically all the useful information that we have about a given image, so we can make changes to policies and introduce new policies without needing to rely on any CLI changes to pass new data.
It also gives us a way of getting local attestations (with the help of containerd) into policy evaluation.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay, I see the idea now. And now that I think about it, I agree it's the right move as well.
Two follow-up thoughts then:
- This won't be a "legacy" thing if it's how we're planning to do local eval in the long-run, so let's not put it in the legacy package.
- Let's try to remove the corresponding mocking logic from go-skill and
scout-cli-plugin
as we are given opportunities to, so that we don't end up holding onto them for some cross-project version combinations.- The packages query mock in this repo will need to remain to support old CLI versions.
- The user query mock in this repo is released and used by
non-root-user-skill
, so we'll need to hold onto it for a while as well. - The packages mock in
scout-cli-plugin
that I linked above is not used by any released skills, to my knowledge. Once this PR is merged, I'll make surepolicy-vulnerabilities
is updated to use the SBOM version, and the mock is removed.
This PR extends the sync-request metadata to include the new
SBOM
field. The Scout CLI will use this field when evaluating a policy locally to send the SBOM to the policy skill. The policy skill will consume it to build the mocks locally.This is piece of work is part of https://www.notion.so/dockerinc/Local-evaluation-Add-support-for-local-attestations-containerd-5a692d03dda94fd1970565113f1c7127?pvs=4.