Authentication Bypass in PraisonAI Call API via Host Header Spoofing (CVE-2026-61435)

hello@craftedsignal.io — Wed, 15 Jul 2026 12:30:37 +0000

CVE-2026-61435 describes a high-severity authentication bypass vulnerability affecting PraisonAI software versions prior to 4.6.78. This flaw specifically impacts the Call API agent invocation endpoints, located in src/praisonai/praisonai/api/agent_invoke.py, when the PRAISONAI_CALL_AUTH=disabled configuration option is enabled. The vulnerability stems from an insecure method of deriving the bind host for a localhost-only safeguard. Instead of using a trusted source, the system relies on the request.url.hostname value, which is populated directly from the client-controlled HTTP Host header. A remote, unauthenticated attacker can exploit this by crafting a request with a spoofed Host: 127.0.0.1 header. This malicious request bypasses the intended localhost-only restriction, enabling the attacker to list all registered agents and invoke them without any authentication, potentially leading to unauthorized access, command execution, or data manipulation depending on the capabilities of the agents.

Attack Chain

An attacker identifies a PraisonAI instance exposed to the network, likely through reconnaissance or internet-wide scanning.
The attacker determines that the target PraisonAI instance is configured with PRAISONAI_CALL_AUTH=disabled.
The attacker crafts an HTTP GET request to the /api/v1/agents endpoint, including a spoofed Host: 127.0.0.1 header.
The PraisonAI server processes the request, mistakenly identifying it as a local request due to the spoofed Host header, and bypasses the authentication check.
The server responds to the attacker with a list of all registered agents on the system, including their agent_id values.
Using the obtained agent_id, the attacker crafts an HTTP POST request to the /api/v1/agents/{agent_id}/invoke endpoint, again including the spoofed Host: 127.0.0.1 header.
The PraisonAI server invokes the specified agent without requiring authentication, allowing the attacker to execute arbitrary actions defined by the agent's functionality.
This results in unauthorized control over PraisonAI agents, leading to potential data exposure, system modification, or further compromise.

Impact

Successful exploitation of CVE-2026-61435 grants an unauthenticated, remote attacker full unauthorized access to list and invoke any registered PraisonAI agent. The impact can range from information disclosure, such as gaining insights into the agents' configurations and capabilities, to significant system compromise. Depending on the functions and permissions of the invoked agents, an attacker could potentially execute arbitrary code, exfiltrate sensitive data, manipulate internal processes, or disrupt services. The broad nature of "agent invocation" implies a high degree of control over the PraisonAI environment and any systems or data the agents can access.

Recommendation

Immediately upgrade PraisonAI to version 4.6.78 or later to patch CVE-2026-61435.
Review your PraisonAI configuration and ensure PRAISONAI_CALL_AUTH is not set to disabled unless absolutely necessary and appropriate compensating controls are in place.
Deploy the Sigma rule "Detects CVE-2026-61435 Exploitation - PraisonAI Authentication Bypass" to your SIEM and investigate any alerts. This rule specifically targets the Host header spoofing behavior.
Monitor webserver logs for HTTP requests to /api/v1/agents or /api/v1/agents/{agent_id}/invoke endpoints that include a Host: 127.0.0.1 header from external IP addresses.

PraisonAI Server-Side Request Forgery via DNS Rebinding and Redirects (CVE-2026-61429)

hello@craftedsignal.io — Sat, 11 Jul 2026 14:23:23 +0000

PraisonAI versions before 1.6.78 are affected by a high-severity server-side request forgery (SSRF) vulnerability, identified as CVE-2026-61429. This flaw resides within the Crawl4AI/Chromium backend component of the application. Attackers can exploit this by crafting malicious URLs that leverage DNS rebinding and HTTP redirects to bypass the application's internal SSRF validation mechanisms. Initially, the crafted URL resolves to a permitted IP address, passing validation checks. However, post-validation, the URL either re-resolves to an internal IP via DNS rebinding or redirects the embedded headless browser to an arbitrary internal service. This allows the attacker to force the PraisonAI application to make requests to internal network resources, access sensitive responses, and potentially exfiltrate critical information such as internal network configurations or security canary values. This vulnerability poses a significant risk for organizations using unpatched PraisonAI installations, as it can lead to internal network reconnaissance and data exfiltration.

Attack Chain

An attacker crafts a malicious URL designed to initially point to an allowed external resource or an IP address that passes PraisonAI's SSRF validation checks.
The attacker initiates a request to the PraisonAI application, instructing its Crawl4AI/Chromium backend to access the crafted URL.
PraisonAI's backend performs its initial SSRF validation against the provided URL, which successfully passes due to the attacker's pre-validation setup (e.g., the URL resolving to a whitelisted IP).
Immediately after validation, the attacker rapidly changes the DNS record for the domain embedded in the crafted URL, making it resolve to an internal IP address or hostname within the target organization's network (DNS rebinding).
Alternatively, the crafted URL is configured to issue an HTTP redirect to an internal IP address or service (e.g., http://192.168.1.100/admin).
The headless Chromium browser in PraisonAI's backend follows the DNS rebind or HTTP redirect, now accessing an internal network resource that would normally be protected from external access.
The application fetches content from the internal service, which can include sensitive data, configuration files, or internal API responses, such as "sensitive canary values."
The attacker then exfiltrates the retrieved internal data from the PraisonAI application's response or logs.

Impact

Successful exploitation of CVE-2026-61429 grants attackers unauthorized access to internal services and potentially sensitive data within the compromised organization's network. This can lead to internal network reconnaissance, allowing attackers to map network infrastructure, identify vulnerable internal systems, and collect confidential information like API keys, credentials, or other internal "canary values" used for security monitoring. The direct consequences can range from data breaches and compliance violations to further exploitation of internal systems, escalating the attack to more severe levels. While no specific victim counts or sectors were provided in the NVD entry, any organization utilizing vulnerable PraisonAI versions is at risk.

Recommendation

Patch CVE-2026-61429 immediately by upgrading PraisonAI installations to version 1.6.78 or later, as specified in the NVD entry.
Implement strict outbound network filtering for the PraisonAI application to prevent unexpected connections to internal or unauthorized external IP addresses.
Monitor DNS queries originating from the PraisonAI application server for sudden changes in resolution for previously whitelisted domains or resolutions to internal IP addresses, indicating potential DNS rebinding attempts.
Review web server logs (category webserver) for HTTP redirects issued by the PraisonAI application to internal network destinations or unusual URLs.

Cpe:2.3:a:praisonai:praisonai:*:*:*:*:*:*:*:* - CraftedSignal Threat Feed

Authentication Bypass in PraisonAI Call API via Host Header Spoofing (CVE-2026-61435)

Attack Chain

Impact

Recommendation

PraisonAI Server-Side Request Forgery via DNS Rebinding and Redirects (CVE-2026-61429)

Attack Chain

Impact

Recommendation

Cpe:2.3:a:praisonai:praisonai:::::::: - CraftedSignal Threat Feed