ASR AI Security Radar
Back to homepage

Recent AI Security Incidents

This archive includes all published incident pages. Page 9 of 18.

Each page is intended to help a security team answer three questions quickly: why the issue is AI-relevant, what part of the workflow may be exposed, and what actions should happen first.

Selection criteria and correction policy are documented in Methodology & Editorial Policy.

AI security incident: Zumba Json Serializer has a potential PHP Object Injection via Unrestricted @type in...

Incident date: February 19, 2026 | Published: February 25, 2026

Description The zumba/json-serializer library allows deserialization of PHP objects from JSON using a special @type field. Prior to version 3.2.

Read details

AI security incident: Dagu affected by unauthenticated RCE via inline DAG spec in default configuration (GH...

Incident date: February 19, 2026 | Published: February 25, 2026

Summary Dagu's default configuration ships with authentication completely disabled. The POST /api/v2/dag-runs endpoint accepts an inline YAML spec and executes its shell commands immediately — no credentials, no token, nothing.

Read details

AI security incident: OpenClaw has a path traversal in apply_patch could write/delete files outside the wor...

Incident date: February 19, 2026 | Published: February 25, 2026

Summary In affected versions, when apply_patch was enabled and the agent ran without filesystem sandbox containment, crafted paths could cause file writes/deletes outside the configured workspace directory.

Read details

AI security incident: Feathers has an origin validation bypass via prefix matching (GHSA-mp4x-c34x-wv3x)

Incident date: February 19, 2026 | Published: February 25, 2026

The origin validation uses startsWith() for comparison, allowing attackers to bypass the check by registering a domain that shares a common prefix with an allowed origin.

Read details

AI security incident: Feathers has an open redirect in OAuth callback enables account takeover (GHSA-ppf9-4...

Incident date: February 19, 2026 | Published: February 25, 2026

Description The redirect query parameter is appended to the base origin without validation, allowing attackers to steal access tokens via URL authority injection.

Read details

AI security incident: Formwork Improperly Managed Privileges in User creation (GHSA-34p4-7w83-35g2)

Incident date: February 19, 2026 | Published: February 25, 2026

Summary The application fails to properly enforce role-based authorization during account creation. Although the system validates that the specified role exists, it does not verify whether the current user has sufficient privileges to...

Read details

AI security incident: Leaf-kit html escaping does not work on characters that are part of extended grapheme...

Incident date: February 19, 2026 | Published: February 25, 2026

Summary htmlEscaped in leaf-kit will only escape html special characters if the extended grapheme clusters match, which allows bypassing escaping by using an extended grapheme cluster containing both the special html character and some...

Read details

AI security incident: Cilium may not enforce host firewall policies when Native Routing, WireGuard and Node...

Incident date: February 19, 2026 | Published: February 25, 2026

Impact [Host Policies](https://docs.cilium.io/en/stable/security/policy/language/#host-policies) will incorrectly permit traffic from Pods on other nodes when all of the following configurations are enabled: * [Native Routing](https://docs.

Read details

AI security incident: CVE-2026-26057 (NVD)

Incident date: February 19, 2026 | Published: February 20, 2026

Skill Scanner is a security scanner for AI Agent Skills that detects prompt injection, data exfiltration, and malicious code patterns.

Read details

AI security incident: Keycloak: Missing Check on Disabled Client for Docker Registry Protocol (GHSA-fjf4-6f...

Incident date: February 19, 2026 | Published: February 25, 2026

A flaw was identified in the Docker v2 authentication endpoint of Keycloak, where tokens continue to be issued even after a Docker registry client has been administratively disabled.

Read details