Isolate registry verifier tests from fork metadata

[xxxxxx-oss]

• This is a small, self-contained fix that does not need prior discussion
• I have read and followed CONTRIBUTING.md and AI_POLICY.md
• I understand and can explain every submitted change and all information in this PR description
• This change is complete, locally validated, and ready for review
• I ran make format, then ran make pre-commit locally and confirmed it passed
• I ran all relevant tests locally and described the broader-suite limitation below
• I have not modified RELEASES.md

Summary

Pin the official repository identity inside the published-registry test fixture so GitHub Actions running in a fork cannot leak its GITHUB_REPOSITORY value into the fixture. This keeps the fixed crates.io and PyPI publisher data internally consistent while leaving production provenance verification unchanged.

Related issues/PRs

None.

Type of change

• Bug fix (non-breaking)
• New feature (non-breaking)
• Improvement (non-breaking)
• Breaking change (impacts existing behavior)
• Documentation update
• Maintenance / chore

Breaking change details (if applicable)

Not applicable.

Documentation

• Documentation changes follow the style guide (not applicable; no documentation changed)
• For PyO3 binding or wrapped Rust doc changes, I ran make py-stubs (not applicable)

Testing

• Affected code paths are already covered by the test suite
• I added/updated tests to cover new or changed logic
• No logic changed (documentation, comments, or metadata only)

Validated locally with:

• bash -n scripts/ci/test-verify-published-registries-crates.bash
• the registry test under both fork and official GITHUB_REPOSITORY values
• file-scoped shfmt, shellcheck, trailing-whitespace, and diff checks
• the remaining release-verification script tests after the unrelated failure described below

Full make test-scripts on macOS stops before this test in the unchanged test-publish-wheels.bash orphan-deletion assertion (Failed orphan deletion left a stale index link). The affected registry test passes independently. Full make format and make pre-commit were not run locally because of local resource constraints; the file-scoped hooks relevant to this Bash-only change passed.