fix: trust bot-applied labels on trusted mirror repos

[Crystora]

Summary

Fixes mirror label_multiplier for trusted first-party repos that apply scoring labels via bots/Apps (which surface in mirror data with actor_association=NULL).

Today the mirror filters labels to actors in MAINTAINER_ASSOCIATIONS (OWNER/MEMBER/COLLABORATOR). Bot-applied labels get NULL actor_association and are silently dropped, so PRs that should resolve to feature (1.5×), bug (1.25×), or refactor (0.5×) collapse to the neutral 1.0×. Concrete example: PR #811 (entrius/gittensor) is labeled refactor by an automation but scores at 1.0× instead of 0.5×.

This PR adds a per-repo opt-in flag rather than weakening the global rule. Repos that haven't earned the trusted_label_pipeline flag keep the existing maintainer-only check unchanged.

This change:

• Adds trusted_label_pipeline: bool = False to RepositoryConfig (defaults to False — fully backwards compatible).
• Threads repo_config into _resolve_maintainer_set_label. When trusted_label_pipeline=True, the actor check is bypassed so any actor (including bots with NULL actor_association) can apply a scoring label; otherwise the existing MAINTAINER_ASSOCIATIONS rule is enforced.
• Enables trusted_label_pipeline=true for entrius/allways, entrius/allways-ui, entrius/das-github-mirror, entrius/gittensor, entrius/gittensor-ui.
• Adds entrius/das-github-mirror to master_repositories.json (weight: 0.2, mirror_enabled: true, trusted_label_pipeline: true).
• Tests:
  • 12-case parametrized truth table on _resolve_maintainer_set_label covering OWNER/MEMBER/COLLABORATOR/CONTRIBUTOR/NONE/NULL actors across trusted_label_pipeline true/false.
  • 4 integration tests in TestPrMultipliers covering bot/maintainer × trusted/untrusted label resolution end-to-end.
  • Loader-level tests asserting trusted_label_pipeline parses correctly from JSON and that every live config entry has the field as a bool.

Related Issues

Fixes #911

Type of Change

• Bug fix
• New feature
• Refactor
• Documentation
• Other

Testing

• Tests added/updated
• Manually tested

Commands run (all green):

• uv run pre-commit run --all-files — ruff lint + format
• uv run pre-commit run --all-files --hook-stage pre-push — pyright + pytest
• uv run pyright — 0 errors, 0 warnings
• uv run pytest tests/ — 723 passed
• uv run pytest tests/validator/oss_contributions/mirror/test_scoring.py tests/validator/test_load_weights.py -v — 98 passed

Checklist

• Code follows project style guidelines
• Self-review completed
• Changes are documented (if applicable)

[anderdc]

Closing in favor of #913 per the maintainer-judgment criterion in #911 ('most elegant, most optimal, most forward-thinking'). Same five deliverables, ~30% smaller diff, cleaner commit history.