Forensia CTI Core

Summary

Ingested excerpt (first ~500 chars of normalized text).

When prompts become shells: RCE vulnerabilities in AI agent frameworks New research exposes how prompt injection in AI agent frameworks can lead to remote code execution. Learn how these vulnerabilities work, what’s impacted, and how to secure your agents. The post When prompts become shells: RCE vulnerabilities in AI agent frameworks appeared first on Microsoft Security Blog . In this article A representative case study: Semantic Kernel CVE-2026-26030: In-Memory Vector Sto…

https://www.microsoft.com/en-us/security/blog/2026/05/07/prompts-become-shells-rce-vulnerabilities-ai-agent-frameworks

sha256:3afcf406f08e4296d94b2e0ebfef6c00c37499fa97c29dfa59d87d6c48cff249

What we pulled out

Deterministic extractor (IOC + allowlisted tokens + ATT&CK IDs present in DB).

cve2
md50
url1
ipv40
ipv60
sha10
domain33
sha2560
malware_token0
attack_pattern2
threat_actor_token0

Indicators

Linked with report → mentions → indicator. Values open the indicator workspace.

Type	Value	Link
domain	cmd.exe	Open →
domain	powershell.exe	Open →
domain	github.com	Open →
domain	rundll32.exe	Open →
domain	net.exe	Open →
domain	certutil.exe	Open →
domain	whoami.exe

Malware families

Allowlist token matches only.

None extracted.

Threat actors mentioned

Allowlist mentions — not a formal attribution verdict.

None extracted.

ATT&CK techniques

MITRE IDs referenced in text and present in local technique table.

None linked (import ATT&CK or adjust text).

CONTINUE INVESTIGATION

High-signal pivots without leaving the thread you started in search.

All reports
Browse the report corpus.
Intel graph (first IOC)
Neighborhood from the first linked indicator.

← Back to reports