Summary
YesWiki contains a server-side request forgery vulnerability in a publicly accessible form actor inbox route. An unauthenticated attacker may cause the server to make outbound HTTP requests to internal services, cloud metadata endpoints, or other hosts reachable from the server when ActivityPub is enabled for a Bazar form.
Affected context
The vulnerability occurs in YesWiki deployments using the affected form actor inbox endpoint with public access control. It requires ActivityPub to be enabled on at least one Bazar form, after which the endpoint processes an HTTP Signature header containing a URL before cryptographic verification and URL safety checks are completed.
Exploit preconditions
- Attack surface: Public HTTP POST access to the form actor inbox API route
- Required access: No authentication is generally required when the route is publicly exposed
- Deployment condition: ActivityPub must be enabled on at least one Bazar form
- Trigger point: Processing of the HTTP Signature
keyIdvalue - Why exploitation is possible: The server retrieves the URL supplied in the signature before validating the signature or restricting outbound destinations
Why this happens
The underlying bug is an order-of-operations failure in HTTP Signature processing. The application treats a user-controlled keyId URL as a resource to fetch before completing cryptographic verification and applying destination validation.
This behavior maps to CWE-918: Server-Side Request Forgery, because attacker-controlled input influences requests made from the server's network position. The insecure pattern allows externally supplied URLs to cross a trust boundary without a strict allowlist, scheme restriction, or network-layer control.
Potential impact
- The server may make unauthorized requests to internal services
- Attackers could discover reachable hosts and network services
- Cloud metadata endpoints may be accessed from vulnerable cloud deployments
- Internal administrative interfaces may be exposed through response, timing, or error signals
- Internal systems or credentials could potentially be targeted for further compromise
- Availability may be affected if the server is induced to make excessive outbound requests
General mitigation guidance
Mitigation generally focuses on eliminating attacker control over outbound request destinations and ensuring validation occurs before any network access.
Common defensive patterns include:
- Allowing outbound requests only to explicitly approved hosts and schemes
- Rejecting private, loopback, link-local, multicast, and otherwise restricted address ranges
- Resolving hostnames safely and rechecking the resulting addresses to reduce DNS-rebinding risk
- Performing complete signature verification before retrieving remote key material where the protocol permits
- Applying short timeouts, redirect restrictions, response-size limits, and controlled egress networking
- Disabling unnecessary public exposure of the affected route or ActivityPub functionality
- Running the application with least-privilege network access and without access to sensitive metadata services
How teams detect issues like this
High-level detection can combine code analysis, runtime monitoring, and deployment policy enforcement.
Useful approaches include:
- Static analysis for HTTP clients receiving URLs derived from request headers or parameters
- Code review checks for network requests made before authentication, signature verification, or authorization
- Rules identifying unrestricted URL schemes, redirects, hostname resolution, or private-address access
- Egress firewall policies that alert on application traffic to internal ranges and cloud metadata addresses
- Web and application logs correlating public API requests with unusual outbound destinations, timing patterns, or repeated connection failures
- Automated security review tools that flag SSRF-prone data flows and missing network-boundary controls

