fix(ci): restore the undefined warningMessage in spam-detection workflow (#187)

The github-script step references warningMessage in createComment, but its
definition was removed in 9ac0c27 while the usage was kept. When a comment
tripped the threshold, the step threw 'ReferenceError: warningMessage is not
defined', which the surrounding try/catch swallowed — so the scam-warning
comment was never posted (only the label was applied). Restore the message,
built from the variables already in scope (author, foundPatterns,
suspiciousLinks).

.github/workflows/spam-detection.yml

@@ -92,6 +92,11 @@ jobs:
               
               // Trigger on either multiple patterns or high spam score
               if (foundPatterns.length > 2 || spamScore >= 3) {
+                const warningMessage =
+                  `⚠️ Potential scam detected in comment by ${author}:\n` +
+                  `- Suspicious patterns found: ${foundPatterns.join(', ')}\n` +
+                  `${suspiciousLinks ? '- Contains external links\n' : ''}` +
+                  `\n@${context.repo.owner} Please review this comment.`;
                 try {
                   // Create a warning comment
                   await github.rest.issues.createComment({