Unauthenticated Remote Code Execution via HTTP Plugin Loading

Bifrost HTTP transport before 2.0.0 accepts an enabled custom plugin whose path is an HTTP URL through unauthenticated POST /api/plugins when management authentication is disabled (the default, governance.auth_config.is_enabled=false). The shared-object loader treats an http-prefixed path as a download URL, writes the body to a temporary .so, and passes it to Go's plugin.Open. After a successful open, optional Init runs immediately with the supplied config as the Bifrost process user. On documented dynamically linked builds (DYNAMIC=1 / no static-link flags), which the vendor requires for custom Go plugins, plugin.Open is expected to succeed and this is unauthenticated remote code execution. On the published statically linked Docker image, plugin.Open fails with Dynamic loading not supported, so that build class is only server-side request forgery. Attack complexity is High because the attacker cannot force RCE on the default static image and a loadable plugin must match the host Go ve

Casky was already ahead

This CVE exploits attack patterns that Casky's 443matched skills already investigate — long before this vulnerability was disclosed. Claude's reasoning model maps these techniques to MITRE ATT&CK, so practitioners who ran these skills have already seen the threat behaviour in their findings.

Analysis

Bifrost HTTP transport versions prior to 2.0.0 contain a critical remote code execution vulnerability that allows unauthenticated attackers to execute arbitrary code on vulnerable systems. The flaw exists in the plugin management endpoint (/api/plugins) which, when management authentication is disabled (the default configuration), accepts POST requests to load custom plugins from HTTP URLs. The vulnerability chains multiple weaknesses: lack of authentication (CWE-306), improper access control (CWE-284), and code injection via unsafe plugin loading (CWE-94). Any organization running Bifrost with default settings—where governance.auth_config.is_enabled=false—is immediately at risk, as attackers can remotely download and execute malicious shared objects (.so files) with the privileges of the Bifrost process.

Casky.ai's 443 mapped security skills leverage Claude's extended reasoning to detect attack patterns across all four phases of this exploitation chain. Practitioners using Casky would identify reconnaissance and resource development activities (TA0001, TA0043) as attackers probe for Bifrost instances and prepare malicious payloads. Detection would escalate when exploitation techniques (TA0004) manifest—specifically unauthorized plugin upload attempts and HTTP requests to /api/plugins without valid credentials. The persistence and privilege escalation phase (TA0006) becomes visible through monitoring of plugin initialization routines and unexpected process execution under the Bifrost service account. Casky's skill-based analysis would surface findings linking unauthenticated POST requests, temporary .so file creation in system directories, and suspicious child process spawning—enabling practitioners to correlate these indicators and respond before arbitrary code reaches execution.

Live Threat IntelligenceComing soon

Casky Risk Score

—

EPSS Probability

—

Shodan Exposure

—

GreyNoise Signal

—

Composite risk scoring from EPSS, CISA KEV, Shodan, and GreyNoise — 21 security APIs correlated into a single Casky Risk Score. Coming in Casky Pro. Join early access →

443 Casky skills cover this attack pattern

These skills use Claude AI's reasoning model to surface findings in the same attack categories as CVE-2026-86242.

Abuse of Public-Facing API: Mobile API Abuse

mobile security · low

Run

abusing-dpapi-for-credential-access

red teaming · high

Run

abusing-shadow-credentials-for-privesc

red teaming · high

Run

MITRE ATT&CK Techniques

TA0001 TA0043 TA0004 TA0006

Investigate the attack patterns behind CVE-2026-86242

Casky has 443 skills that investigate the attack patterns behind CVE-2026-86242. Run one and get CVSS-scored findings in 3 minutes.

Run the skill that detects this →

X Instagram LinkedIn