Skip to content

Make AI policy rules mechanically enforceable#33

Open
damusix wants to merge 1 commit into
hapijs:masterfrom
damusix:docs/amend-ai-policy
Open

Make AI policy rules mechanically enforceable#33
damusix wants to merge 1 commit into
hapijs:masterfrom
damusix:docs/amend-ai-policy

Conversation

@damusix

@damusix damusix commented Jun 12, 2026

Copy link
Copy Markdown
Contributor

The current policy boils down to "be transparent and don't waste our
time". That standard is open to interpretation, which means reviewers
end up litigating quality case by case.

This revision retains the core expectations (defend-every-line, minimal
diffs, purposeful tests, docs) and makes every rule for outside
contributions mechanically enforceable:

  • Disclosure is mandatory — tool + extent in the PR description,
    Assisted-by: trailer in commits. Suspected-but-undisclosed use closes
    the PR, no argument required.
  • AI-assisted PRs require an accepted issue. Drive-by PRs are closed
    unreviewed. AI-assisted issues stay welcome (human in the loop).
  • Repeat violations escalate to a ban.
  • Maintainer AI use is stated openly and goes through the same peer
    review as everything else.

Structure follows RustPython's policy
(derived from Ghostty's). The accepted-issue gate is the substantive
change for reviewer burden: maintainers only review AI-assisted PRs for
problems the team has already accepted as real.

"Don't waste our time" left too much to interpretation. Keep the
standing expectations; add rules a maintainer can apply without
debate: mandatory Assisted-by disclosure, accepted-issue gate for
AI-assisted PRs, ban on repeat violations. Maintainer AI use is
now stated openly. Follows the RustPython/Ghostty model.
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