Summary
CoreWorxLab CAAL up to version 1.6.0 contains a server-side request forgery (SSRF) weakness in the test-hass endpoint implemented in src/caal/webhooks.py.
By abusing this endpoint, a remote attacker can coerce the server into making arbitrary outbound HTTP requests, potentially reaching internal services that are not directly exposed to the internet.
Affected context
The issue arises within the test-hass webhook endpoint in CoreWorxLab CAAL, which appears to accept user-controlled URLs or destinations and then performs outbound HTTP requests on behalf of the caller.
Deployments where the CAAL server can reach internal networks, cloud metadata services, or other sensitive HTTP endpoints are more exposed.
The risk increases when network egress from the CAAL host is unrestricted and when internal services rely solely on network location rather than strong authentication.
Exploit preconditions
- Attack surface: Publicly reachable test-hass endpoint in
src/caal/webhooks.py. - Required access: Ability to send HTTP requests to the vulnerable endpoint (typically unauthenticated or low-auth access, depending on deployment).
- Input control: Attacker can influence the URL, host, or path used by the server when performing outbound requests.
- Trigger point: Invocation of the test-hass endpoint that processes attacker-supplied parameters and issues a server-side HTTP request.
- Why exploitation is possible: The endpoint forwards user-controlled input directly into HTTP client logic without sufficient validation, allowlisting, or network-level restrictions, enabling requests to arbitrary or internal targets.
Why this happens
This vulnerability is a classic case of CWE-918: Server-Side Request Forgery (SSRF).
The webhooks implementation for the test-hass endpoint appears to use user-provided values to construct outbound HTTP requests without proper validation or restriction of target addresses, protocols, or paths.
By trusting client-supplied URLs, the server is effectively repurposed as an HTTP proxy that can reach locations inaccessible to the attacker, such as internal services, loopback interfaces, or cloud infrastructure endpoints.
The core design flaw is the lack of strict output-boundary checks and allowlisting on server-initiated network calls that are influenced by user input.
Potential impact
- May allow probing and enumeration of internal network services and ports reachable from the CAAL server.
- Could enable unauthorized access to internal HTTP APIs not directly exposed to the internet.
- Can potentially be used to access cloud instance metadata services, leading to credential or token exposure.
- Might facilitate bypassing IP-based access controls and firewalls by leveraging the CAAL server as a stepping stone.
- Could assist in mapping internal infrastructure as part of a broader intrusion campaign.
General mitigation guidance
Effective mitigation focuses on constraining where the application can connect and reducing trust in user-controlled URLs.
Common defensive patterns include:
- Implementing strict allowlists of permitted target hosts, IP ranges, and schemes for any server-initiated HTTP requests.
- Rejecting or normalizing user input that directly influences network destinations, especially private IP ranges, loopback, link-local, and cloud metadata addresses.
- Introducing an application-layer proxy or service abstraction that exposes only vetted destinations instead of raw URL forwarding.
- Applying least-privilege network egress rules at the host, container, or cluster level to block outbound traffic to sensitive internal segments where it is not required.
- Avoiding passing full URLs from clients; instead, allow selection from predefined options managed in server-side configuration.
How teams detect issues like this
Detection typically combines code-level analysis and runtime monitoring:
- Static analysis: Search for HTTP client calls (e.g., requests to
requests,urllib, or similar libraries) where the target URL or host is derived directly from request parameters or headers, particularly insrc/caal/webhooks.pyand related handlers. - Code review signals: Look for endpoints labeled as "test," "webhook," or "callback" that accept arbitrary URLs, hostnames, or IPs and pass them to networking APIs without normalization or allowlisting.
- Security testing: Use dynamic application security testing (DAST) or manual penetration testing to send URLs pointing to internal ranges (e.g.,
127.0.0.1,10.0.0.0/8, metadata IPs) and observe whether the server responds with data from those locations. - Operational monitoring: Inspect outbound traffic logs from the CAAL server for unusual destinations such as internal subnets, loopback addresses, or cloud metadata endpoints, and alert on anomalous patterns.

