Aculect AI Companion EUVD-2026-91663

Exploitation requires a victim to click an attacker-crafted link that targets the plugin's unvalidated redirect/forward parameter (UI:R in the vector), so it cannot be triggered silently or server-side. No authentication is required (PR:N) and the flaw is reachable over the network against default installations of the Aculect AI Companion plugin (versions through 0.8.1). The practical limiting factors are the mandatory user interaction, the low-value impact (redirection/phishing only, no data access or code execution), and the apparently small deployment footprint of this specific plugin.

This is a moderate-severity open redirect (CWE-601) with a CVSS 3.1 base score of 4.7, consistent with its vector (AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:N/A:N): network-reachable and unauthenticated but requiring user interaction, with only low confidentiality impact and no integrity or availability impact. No EPSS score, SSVC decision, KEV listing, or POC indicator was supplied in the input, so there is no public exploit identified at time of analysis and no evidence of active exploitation. The signals are internally consistent - a low-impact phishing-enabler flaw, not a code-execution bug - so this reads as a genuinely low-priority issue rather than a high-CVSS-but-low-real-risk mismatch. The dominant risk driver is not technical impact but the social-engineering value of abusing a trusted plugin domain to redirect victims; the UI:R requirement (victim must follow a crafted link) and the niche footprint of the affected plugin further constrain real-world risk. Key missing data: no EPSS probability, no KEV/POC confirmation, and no vendor advisory beyond the Patchstack database entry.

An attacker crafts a URL pointing at a vulnerable Aculect AI Companion endpoint that embeds an attacker-controlled redirect target, then distributes it via email or social media. A victim who trusts the legitimate WordPress site's domain clicks the link and is silently forwarded to an attacker-controlled phishing page. No public exploit code was identified at time of analysis, but the crafted-link technique is trivial given AC:L and no authentication requirement.