Summary
An authenticated Server-Side Request Forgery (SSRF) in Koillection v1.8.0's custom scraper subsystem lets a logged-in attacker supply crafted URLs that the server will fetch on their behalf.
When the scraper is abused to contact internal endpoints, an attacker can scan internal networks or services that are not directly exposed to the internet.
Affected context
The vulnerability occurs in Koillection deployments that expose the custom scraper subsystem to authenticated users and allow it to fetch remote URLs.
It becomes relevant when the application server has network access to internal services (e.g., other containers, internal APIs, metadata endpoints) that are not otherwise reachable from the attacker's network.
The issue is more serious in environments where Koillection runs inside a trusted network segment or data center with access to sensitive internal infrastructure.
Exploit preconditions
- Attack surface: HTTP interface of the Koillection custom scraper feature.
- Required access: Valid authenticated user account with permission to use the scraper functionality.
- Trigger point: Submitting or configuring a crafted URL that the scraper will request server-side.
- Why exploitation is possible: The server-side HTTP client does not sufficiently restrict target addresses or schemes, allowing requests to internal or otherwise sensitive network locations based on user-controlled input.
Why this happens
This is an instance of Server-Side Request Forgery mapped to CWE-918, where a server performs outbound HTTP requests to attacker-specified destinations.
The custom scraper subsystem trusts user-controlled URLs and forwards them directly to an internal-capable HTTP client without robust validation or network egress controls.
In abstract terms, the bug class arises when:
- User input is treated as a fully trusted URL or endpoint.
- The server has broader network reach than the user.
- There are no effective checks to constrain or normalize destinations (such as blocking private address ranges, local interfaces, or sensitive metadata endpoints).
Potential impact
- An attacker may scan internal IP ranges and ports accessible from the Koillection host.
- An attacker could probe internal HTTP services and infer their presence or behavior based on responses.
- An attacker can potentially reach management interfaces, internal APIs, or cloud metadata services that are not exposed externally.
- An attacker may leverage SSRF as a stepping stone towards further vulnerabilities (e.g., exploiting internal services, accessing configuration data, or performing lateral movement).
General mitigation guidance
Effective mitigation focuses on constraining where server-side HTTP clients can connect and reducing trust in user-supplied URLs:
- Apply strict allowlists for outbound destinations or domains that the scraper may contact.
- Block requests to private, loopback, link-local, and metadata-related IP ranges at the application logic and/or network layer.
- Normalize and re-validate parsed URLs (including redirects) before performing a request, enforcing scheme and host constraints.
- Run the application with network segmentation and least-privilege network access so it cannot freely reach sensitive internal services.
- Prefer design patterns where user-supplied URLs are fetched by the client (browser) when feasible, instead of by privileged backend services.
How teams detect issues like this
Detection can combine code-centric and operational approaches:
- Use static analysis or code review to flag patterns where user-controlled URL parameters are passed directly into HTTP clients or scraping utilities.
- Check for missing destination validation logic, absence of IP range checks, and lack of normalization around redirects or DNS resolution.
- Monitor server logs and outbound network traffic for unusual patterns, such as requests to internal IP ranges, localhost, or cloud metadata endpoints originating from the scraper component.
- Employ egress filtering or proxy-based policies that log and restrict outbound HTTP calls from the Koillection server to approved destinations.

