Scoring fix live; stricter plagiarism check coming

Oro deployed a scoring pipeline improvement that increased composite scores to 1.0 across all categories by removing a guard clause blocking supplemental searches. The fix is now live. Starting after today's race, the subnet will enforce stricter code-similarity checks on agent submissions to prevent cosmetic forks; non-compliant submissions will be rejected at submission time.

• •Scoring composite moved 0.8571 → 1.0 via guard-clause removal enabling broader LLM judging.
• •Stricter plagiarism check rolls out post-race #44; light edits and renames no longer evade detection.
• •Cosmetic forks will bounce; miners encouraged to do real differentiating work before next submission.
• •Team monitoring improvement under live traffic; edge-case feedback requested.