Content sniffing for uploads (don't trust the header)

8164 views

                  package api

import (
  "io"
  "net/http"
)

func DetectType(f io.Reader) (string, []byte, error) {
  buf := make([]byte, 512)
  n, err := io.ReadFull(f, buf)
  if err != nil && err != io.EOF && err != io.ErrUnexpectedEOF {
    return "", nil, err
  }
  buf = buf[:n]
  return http.DetectContentType(buf), buf, nil
}

Clients can lie about Content-Type, so for uploads I prefer sniffing the first bytes with http.DetectContentType. The safe flow is: open the multipart file, read a small prefix, detect type, then rewind (or re-open) before writing to disk or object storage. This keeps you from accepting “image/png” that is actually an executable, and it reduces surprises in downstream processing. I also enforce a maximum size at the HTTP layer with MaxBytesReader and validate file extensions only as a secondary signal. In production, I keep an allowlist of MIME types per upload endpoint (avatars vs PDFs vs videos) and I log rejections with request IDs. This snippet demonstrates the minimal sniffing step that makes upload validation meaningfully safer.

Leah Thompson

More from Leah Thompson

Content sniffing for uploads (don't trust the header)

0 Comments

More from Leah Thompson