OK, comments should work now

action.js

@@ -47,13 +47,22 @@ for (const image of images) {
 
 // Prepare reply body.
 const payload = github.context.payload;
-const titleAndBody = payload.issue.title + '\n\n' + payload.issue.body;
+const isComment = payload.comment !== undefined;
+let titleAndBody;
+let author;
+if (isComment) {
+  titleAndBody = payload.comment.body;
+  author = payload.comment.user.login;
+} else {
+  titleAndBody = payload.issue.title + '\n\n' + payload.issue.body;
+  author = payload.issue.user.login;
+}
 const quoteArray = [];
 for (const line of titleAndBody.split('\n')) {
   // Remove some speical chars to remove at mention spam possibilities.
   quoteArray.push('> ' + line.replace(/[@#]/g, ""));
 }
-const replyBody = `Hi ${payload.issue.user.login},
+const replyBody = `Hi ${author},
 
 ${quoteArray.join('\n').substring(0,40000)}
 
@@ -61,9 +70,10 @@ ${full_images.join('\n\n')}
 `;
 
 // Label handling.
-const labels = new Set(payload.issue.labels.map(label => label.name));
-const newLabels = new Set();
-const shabiWords = [
+if (!isComment) {
+  const labels = new Set(payload.issue.labels.map(label => label.name));
+  const newLabels = new Set();
+  const shabiWords = [
     'shabi',
     'shadiao',
     '傻',
@@ -71,38 +81,38 @@ const shabiWords = [
     '智障',
     '啥b',
     'stupid',
-];
-for (const word of shabiWords) {
+  ];
+  for (const word of shabiWords) {
     if (new RegExp(word, 'i').test(titleAndBody)) {
       newLabels.add('you-are-stupid-argument');
       break;
     }
-}
-const fuckMotherWords = [
+  }
+  const fuckMotherWords = [
     'cnm',
     '操你妈',
     'fuck.*\\b(mom|mum|mother)\\b',
     '尼玛',
     '去你吗',
-]
-for (const word of fuckMotherWords) {
+  ]
+  for (const word of fuckMotherWords) {
     if (new RegExp(word, 'i').test(titleAndBody)) {
       newLabels.add('fuck-your-mother-argument');
       break;
     }
-}
-const motherDiedWords = [
+  }
+  const motherDiedWords = [
     'nmsl',
     '你妈死',
     '司马',
-]
-for (const word of motherDiedWords) {
+  ]
+  for (const word of motherDiedWords) {
     if (new RegExp(word, 'i').test(titleAndBody)) {
       newLabels.add('your-mother-died-argument');
       break;
     }
-}
-const shitpostWords = [
+  }
+  const shitpostWords = [
     'fuck',
     'shit',
     'bitch',
@@ -110,19 +120,20 @@ const shitpostWords = [
     '婊子',
     '恶心',
     '操你',
-];
-for (const word of shitpostWords) {
+  ];
+  for (const word of shitpostWords) {
     if (new RegExp(word, 'i').test(titleAndBody)) {
       newLabels.add('shitpost');
       break;
     }
-}
-if (newLabels.size > 0) {
+  }
+  if (newLabels.size > 0) {
     newLabels.add('shitpost');
     if (labels.has('not-shitpost')) {
       labels.delete('not-shitpost');
       newLabels.add('op-does-not-know-what-shit-is');
     }
+  }
 }
 
 // Make the request.
@@ -135,12 +146,15 @@ try {
     issue_number: payload.issue.number,
     body: replyBody,
   });
+  if (!isComment) {
+    // Update labels.
     await octokit.issues.update({
       owner: 'cirosantilli',
       repo: payload.repository.name,
       issue_number: payload.issue.number,
       labels: Array.from([...labels, ...newLabels])
     });
+  }
 } catch (error) {
   core.setFailed(error.message);
 }