Securing non-human identities: automated revocation, OAuth, and scoped permissions

2026-04-15T19:24:09Zed6f39f501e4fd26c06c3b6540740e9db94d07fd5e77960df908adda67946ba3
ACMEAI-securityCloudflare-Radar','radar','CGNAT','Workers-VPC','MCP','log-explrDDoSLLMPingoraRFC-9728account-abuse-protectionai-discoveryapi-securityapi-tokenscertificate-validationclient-side-securitycloudflaredeveloper-securityfraud-preventionleast-privilegemanaged-oauthoauthpost-quantumpost-quantum-cryptographyrequest-smugglingresource-scoped-permissionsscannable-tokensvulnerability-scanner

What happened

Cloudflare published a batch of security updates and product launches focusing on developer and enterprise identity safety, AI-aware protections, and infrastructure hardening. Key highlights include scannable API tokens and resource-scoped permissions (GA) for least-privilege access, Managed OAuth for Access (RFC 9728) to support agent-safe authentication, expanded client-side and AI app security, a new API/Web vulnerability scanner, fixes for request-smuggling in Pingora OSS (Pingora 0.8.0), and an explanation/mitigation for an ACME validation flaw. The posts also cover operational telemetry—

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
cloudflare_security_blog
Record identifier
ed6f39f501e4fd26c06c3b6540740e9db94d07fd5e77960df908adda67946ba3
Enrichment time
2026-04-15T19:24:09Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.