// AI Platform Drift

Cloudflare WAF flags non-multipart Content-Type with boundary parameter

Cloudflare developer changelog · published 2026-11-14 18:05 UTC
WORTH KNOWING VERIFY_ONLY SECURITYAPI
WHAT CHANGEDEffective 28 October 2026, Cloudflare's malicious uploads detection sets cf.waf.content_scan.has_failed to true for requests with a Content-Type boundary parameter but non-multipart type (example: application/json; boundary=example). The body of flagged requests is not scanned.
WHY IT MATTERSRequests with this header combination bypass content scanning entirely. If your WAF rules depend on content_scan results or assume all uploads are scanned, these requests will not match expected security policies.
WHAT TO DOCheck your application's Content-Type headers for boundary parameters on non-multipart requests. Review WAF rules that reference cf.waf.content_scan fields to confirm they handle has_failed=true correctly.
Model-reported confidence 85% (not measured accuracy) Primary source ↗ written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## WAF - Malicious uploads detection flags non-multipart Content-Type headers with a boundary
+https://developers.cloudflare.com/changelog/post/2026-10-28-content-scan-multipart-content-type/
+Wed, 28 Oct 2026 00:00:00 GMT
+
+The malicious uploads detection now sets cf.waf.content_scan.has_failed to true for requests whose Content-Type header has a boundary parameter but type is not multipart/*, for example application/json; boundary=example. The body of these requests is not scanned.
+For more information, refer to Multipart requests.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
triage verdictMATERIAL at 0.92, needed 0.5
triage reasoningWAF behavior change with future effective date alters how malicious upload detection flags and processes requests.
signalsfuture date (2026-10-28) · WAF behavior change documented · new flag condition: cf.waf.content_scan.has_failed set to true · scanning behavior change: body not scanned for flagged requests · affects request handling and security posture
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)