Skip to content

docs: SignPath declined for now — record both remaining routes honestly - #357

Merged
zcsizmadia merged 1 commit into
mainfrom
docs/signpath-rejected
Sep 16, 2026
Merged

zcsizmadia merged 1 commit into
mainfrom
docs/signpath-rejected

Conversation

@zcsizmadia

Copy link
Copy Markdown
Collaborator

The SignPath Foundation's free OSS programme declined the application — it is for projects with an established user base, and Skrog does not have one yet. They invited a reapplication as visibility grows.

Five documents claimed signing was applied for and imminent, and docs/code-signing.md credited SignPath as though they were already sponsoring us. All of that is now false, and it is the kind of false that matters: someone deciding whether to trust an unsigned binary should not be told a signature is on its way when nobody has promised one.

The honest position: two routes, neither closed

free the Foundation, deferred until the project is bigger — and gated on exactly the thing signing would help with, since winget builds reputation fastest and wants a signed installer
paid a subscription, which needs nobody's approval and could happen at any time

What docs/code-signing.md keeps, and what it drops

Keeps the roles, the CI-only signing rule and the threat model. They were written to meet the Foundation's requirements, but they are good practice whoever issues a certificate, and having them settled beats inventing them in a hurry the week a certificate arrives.

Drops the attribution block crediting SignPath for providing free signing. They are not sponsoring this project; leaving it would be misleading.

Adds the part that is actually the decision in #77, because the cheaper option is not automatically the better one:

The Foundation issues the certificate to the project and forbids commercial dual-licensing of any component, and can pause or revoke. A purchased certificate costs money but nobody can withdraw it over a licence change.

So the free route keeps a door open that the paid route closes — and vice versa.

Elsewhere

  • ROADMAP — signing lifted out of the weekend plan. It is a decision, not a scheduled item, so a date next to it was fiction.
  • PLAN — the SmartScreen risk is raised to high, with the circularity stated rather than assumed away: the free route is gated on adoption, winget would drive adoption, winget wants a signed installer.
  • README, install guide, RELEASING — no promised date; both routes named; #77 is where the choice lives.

Docs only, no behaviour change. The thing every release does carry — SLSA provenance and a cosign-signed SHA256SUMS — is unchanged and still stated as not a substitute for signing.

The SignPath Foundation's free OSS programme declined the application: it is
for projects with an established user base, and Skrog does not have one yet.
They invited a reapplication as visibility grows.

Five documents claimed signing was applied for and imminent, and one of them
credited SignPath as though they were already sponsoring us. All of that is now
false, and it is the kind of false that matters: a reader deciding whether to
trust an unsigned binary should not be told a signature is on its way when
nobody has promised one.

The honest position is two open routes, neither closed:

  free   the Foundation, deferred until the project is bigger -- and gated on
         exactly the thing signing would help with, since winget builds
         reputation fastest and wants a signed installer
  paid   a subscription, which needs nobody's approval and could happen at
         any time

docs/code-signing.md keeps the roles, the CI-only rule and the threat model.
They were written to meet the Foundation's requirements but they are good
practice whoever issues a certificate, and having them settled beats inventing
them in a hurry later. The attribution block crediting SignPath is removed --
they are not sponsoring this project, and saying so would be misleading.

Also written down, because it is the actual decision in #77 and the cheaper
option is not automatically the better one: the Foundation issues the
certificate TO THE PROJECT and forbids commercial dual-licensing of any
component, and can revoke. A purchased certificate costs money but nobody can
withdraw it over a licence change. The free route keeps a door open that the
paid route closes, and vice versa.

Signing is lifted out of the ROADMAP's weekend plan -- it is a decision, not a
scheduled item -- and PLAN's SmartScreen risk is raised to high, since the
circularity is now explicit rather than assumed away.
@zcsizmadia
zcsizmadia merged commit 8067a4f into main Sep 16, 2026
4 checks passed
@zcsizmadia
zcsizmadia deleted the docs/signpath-rejected branch September 16, 2026 12:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant