Description

A vulnerability was determined in adenot mcp-google-search up to 0.3.1. Impacted is an unknown function of the file src/index.ts of the component read_webpage. Executing a manipulation of the argument url can lead to server-side request forgery. The attack is restricted to local execution. This patch is called f071d491b685011ca04e8ab8d586fc65f86bcee1. It is advisable to implement a patch to correct this issue.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "5.3",
  "pubDate": "2026-08-09T06:20:07.843Z",
  "pubdate": "2026-08-09T06:20:07.843Z",
  "executiveSummary": "An SSRF vulnerability exists in adenot mcp-google-search up to 0.3.1 within the read_webpage component located in src/index.ts. The flaw allows an attacker to manipulate the url argument, resulting in server-side request forgery.\nThe vulnerability impacts the integrity and confidentiality of the host environment where the affected component executes. By abusing the URL parsing and fetching mechanisms, unauthorized requests can be dispatched from the hosting server.\nExploitation of this vulnerability is restricted to local execution, requiring local access or local context to manipulate the input arguments passed to the vulnerable function.\nThe risk implication involves potential unauthorized interaction with internal network resources or local services that are otherwise inaccessible from external perimeters.\nMitigation requires applying the specific patch identified by commit hash f071d491b685011ca04e8ab8d586fc65f86bcee1 to properly validate and restrict the URL inputs.",
  "technicalDetails": "The vulnerability resides in the read_webpage component implemented in the src/index.ts file of the adenot mcp-google-search package across versions up to 0.3.1.\nThe root cause stems from insufficient validation, sanitization, and restriction of the url argument supplied to the webpage retrieval functionality.\nThe attack vector involves local execution, where an actor with local access or control over the execution parameters manipulates the target URL input passed into the underlying HTTP request mechanism.\nDuring exploitation, the component processes the attacker-supplied url parameter without adequately verifying whether the destination resolves to internal networks, loopback addresses, or sensitive local endpoints.\nStep-by-step attack flow: First, the attacker identifies or supplies a crafted payload targeting an internal or arbitrary URI via the url argument in the read_webpage function. Second, the vulnerable src/index.ts script processes the input and initiates an outbound network request using the malicious URL. Third, the local server executes the request on behalf of the application, potentially interacting with restricted internal services or metadata endpoints.\nAuthentication and privilege requirements are constrained by local execution parameters, meaning the attacker must be able to invoke the local component or influence the input arguments provided to it.\nThe post-exploitation impact includes unauthorized data exposure, internal service probing, and potential manipulation of internal network endpoints reachable from the host executing the vulnerable code."
}