fix: exclude auto-close event from post-merge edit detection in issue discovery scoring

[spider-yamet]

Summary

Fix a false positive in the post-merge edit detection for issue discovery scoring. When a PR merge auto-closes a linked issue, GitHub sets the issue's updated_at to the close timestamp, which is always slightly after merged_at. The existing check treated this as a post-merge content edit, incorrectly flipping the issue from "solved" to "closed" — reducing the discoverer's solved_count, inflating closed_count, and dropping their credibility.

The fix adds a guard that compares updated_at against closed_at. The penalty now only fires when the issue was genuinely edited after the close event, not when the update was the close event itself.

Related Issues

Fixes #443

Type of Change

• Bug fix
• New feature
• Refactor
• Documentation
• Other (describe below)

Manual testing performed:

• Verified that an issue auto-closed by a merged PR (where updated_at == closed_at) is correctly classified as "solved"
• Verified that an issue with updated_at > closed_at (genuine post-close edit) still triggers the penalty
• Verified that an issue with closed_at = None conservatively triggers the penalty

Checklist

• Code follows project style guidelines
• Self-review completed

[anderdc]

Closing in favor of #434 which replaces the updated_at check entirely with precise body/title edit detection via GraphQL timeline events. That fix covers this case and all other false positives (bot comments, labels, reactions).