romtuck/openclaw
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(media): skip audio files in extractFileBlocks text extraction

Browse Source 
Audio files (especially OGG/Opus from Telegram voice messages) were being
misidentified as text by looksLikeUtf8Text() because OGG headers contain
>85% printable ASCII. This caused guessDelimitedMime() to classify them
as text/tab-separated-values, injecting raw binary into the model context.

Add audio to the skip list alongside image and video in
extractFileBlocks() so audio attachments are routed to the transcription
pipeline instead of being treated as text files.

Fixes #1989
This commit is contained in:
nullrunner 2026-01-29 23:45:40 +01:00
parent 4583f88626
commit 67bbacb3b7
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/media-understanding/apply.ts
View File

@ -216,7 +216,7 @@ async function extractFileBlocks(params: {
}
const forcedTextMime = resolveTextMimeFromName(attachment.path ?? attachment.url ?? "");
const kind = forcedTextMime ? "document" : resolveAttachmentKind(attachment);
if (!forcedTextMime && (kind === "image" || kind === "video")) {
if (!forcedTextMime && (kind === "image" || kind === "audio" || kind === "video")) {
continue;
}
if (!limits.allowUrl && attachment.url && !attachment.path) {