diff --git a/src/github.js b/src/github.js
index e568f56..7f7ae7d 100644
--- a/src/github.js
+++ b/src/github.js
@@ -1,4 +1,5 @@
 import { USER } from './user.js';
+import { btou } from './utils.js';
 
 export const ghApi = async (path, method = 'GET', body = null) => {
   const t = USER.githubToken;
@@ -22,3 +23,25 @@ export const parseGhUrl = u => {
     repo = repoPart.split('@')[0], path = p.slice(2).join('/').replace(/\/$/, '');
   return { owner, repo, branch, path, apiPath: `${owner}/${repo}/contents${path ? '/' + path : ''}` };
 };
+
+// Fetch a file's text content, transparently handling GitHub's 1MB Contents API
+// limit. Files >1MB come back from the Contents API with encoding "none" and an
+// empty `content` field, so we fall back to the Git Blobs API (supports up to 100MB).
+export const ghGetFileContent = async (info, fileName) => {
+  const meta = await ghApi(`${info.apiPath}/${fileName}?ref=${info.branch}`);
+  if (!meta) { console.warn('[Sune] GH file not found:', fileName); return null; }
+  console.log('[Sune] GH file meta:', { name: fileName, size: meta.size, encoding: meta.encoding, hasContent: !!(meta.content && meta.content.trim()), sha: meta.sha });
+  if (meta.content && meta.encoding === 'base64') {
+    try { return btou(meta.content); } catch (e) { console.error('[Sune] decode (contents) failed:', e); }
+  }
+  // Large file path: Contents API omitted the body, retrieve the raw blob by sha.
+  if (meta.sha) {
+    try {
+      const blob = await ghApi(`${info.owner}/${info.repo}/git/blobs/${meta.sha}`);
+      console.log('[Sune] GH blob:', { size: blob?.size, encoding: blob?.encoding, hasContent: !!(blob?.content && blob.content.trim()) });
+      if (blob && blob.content && blob.encoding === 'base64') return btou(blob.content);
+    } catch (e) { console.error('[Sune] blob fetch failed:', e); }
+  }
+  console.warn('[Sune] Could not retrieve content for', fileName);
+  return null;
+};