Host Header Spoofing Bypasses Authentication in PraisonAI

PraisonAI before 4.6.78 contains an authentication bypass in the Call API agent invocation endpoints (src/praisonai/praisonai/api/agent_invoke.py) when PRAISONAI_CALL_AUTH=disabled is configured. The safeguard intended to restrict the disabled-auth opt-out to localhost binding derives the bind host from request.url.hostname, which is taken from the client-controlled HTTP Host header. A remote, unauthenticated attacker who can reach the service over the network can send a spoofed 'Host: 127.0.0.1' header to bypass the localhost-only restriction and list (GET /api/v1/agents) and invoke (POST /api/v1/agents/{agent_id}/invoke) registered agents without authentication.

Casky was already ahead

This CVE exploits attack patterns that Casky's 283matched 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

PraisonAI versions before 4.6.78 contain a critical authentication bypass vulnerability in the Call API agent invocation endpoints. When the PRAISONAI_CALL_AUTH=disabled configuration is set, the service is intended to restrict unauthenticated access to localhost connections only. However, the authentication check incorrectly derives the bind host from the client-controlled HTTP Host header rather than the actual server binding address. This allows remote, unauthenticated attackers to bypass the localhost-only safeguard by simply spoofing the Host header to "127.0.0.1," gaining unauthorized access to sensitive API agent invocation capabilities. Organizations running PraisonAI with disabled authentication are immediately vulnerable to complete compromise of their AI agent functionality, potentially enabling attackers to execute arbitrary agent commands, exfiltrate sensitive data, or manipulate system operations.

Casky.ai's 283 mapped security skills leverage Claude AI with extended reasoning to detect the attack patterns behind this vulnerability across MITRE ATT&CK techniques TA0004 (Privilege Escalation) and TA0006 (Credential Access). Practitioners using Casky would identify suspicious findings including: anomalous HTTP requests with spoofed Host headers targeting agent invocation endpoints, successful API authentications originating from non-localhost source IPs despite disabled-auth configuration, and patterns of unauthenticated access to /api/agent_invoke paths. The platform's skills would flag authentication enforcement bypass attempts by correlating Host header values against actual server binding addresses, detecting when remote IPs successfully authenticate via localhost header spoofing. Extended reasoning across credential access techniques would reveal lateral movement patterns where attackers abuse the bypassed authentication to invoke agents and escalate privileges within connected systems or environments.

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 →

283 Casky skills cover this attack pattern

These skills use Claude AI's reasoning model to surface findings in the same attack categories as CVE-2026-61435.

abusing-dpapi-for-credential-access

red teaming · high

Run

abusing-shadow-credentials-for-privesc

red teaming · high

Run

Access with Stolen Session Cookie

identity access management · low

Run

MITRE ATT&CK Techniques

TA0004 TA0006

Investigate the attack patterns behind CVE-2026-61435

Casky has 283 skills that investigate the attack patterns behind CVE-2026-61435. Run one and get CVSS-scored findings in 3 minutes.

Run the skill that detects this →

X Instagram LinkedIn