.github/scripts/bot-next-issue-recommendation.js (1)

55-65: ⚠️ Potential issue | 🔴 Critical

Critical: labelSet.has('Good First Issue') will never match — recommendation bot is broken for GFIs.

Line 55 lowercases every label name before inserting into labelSet, so the Set only ever contains entries like 'good first issue'. Set.prototype.has uses SameValueZero (case-sensitive) equality, so labelSet.has('Good First Issue') always returns false. As a result:

• difficultyLevels.goodFirstIssue is permanently false.
• The early return at line 74 silently skips GFI-only issues (those without a beginner label).
• At line 97, completedLabel is mislabeled as 'Beginner' even when the issue is a GFI.

This regresses the assertion in the PR description — "Preserved existing lowercase comparisons that use .toLowerCase() to avoid changing logic." The comparison side here was changed and breaks the gating.

🐛 Proposed fix — keep the comparison case-insensitive (consistent with `bot-gfi-assign-on-comment.js`)

     // Determine issue difficulty level
     const difficultyLevels = {
       beginner: labelSet.has('beginner'),
-      goodFirstIssue: labelSet.has('Good First Issue'),
+      goodFirstIssue: labelSet.has('good first issue'),
       intermediate: labelSet.has('intermediate'),
       advanced: labelSet.has('advanced'),
     };

Alternatively, lift 'Good First Issue' into a top-level constant and lowercase it at the comparison site — that would match the pattern used in bot-gfi-assign-on-comment.js (GOOD_FIRST_ISSUE_LABEL.toLowerCase()) and serve as the single source of truth requested in issue #2178.

As per coding guidelines: "Use top-level constants for configuration — avoid hardcoded values scattered through the script."