CVE WatchAug 6, 2026

CVE-2026-55391

datamodel-code-generator generates Pydantic v2 models, dataclasses, TypedDict, and msgspec.Struct from OpenAPI, JSON Schema, GraphQL, Avro,

CVSS

7.5

High

EPSS

0.2%

p15

KEV

—

Exploit Today

0-100

Published: Jul 28, 2026 · Last modified: Aug 6, 2026 · CWE-350 · CWE-367 · CWE-918

EPSS · 30d

0.2%EPSS · 30 days0.2%

2026-08-182026-09-15

Technical description

datamodel-code-generator generates Pydantic v2 models, dataclasses, TypedDict, and msgspec.Struct from OpenAPI, JSON Schema, GraphQL, Avro, Protobuf, and raw JSON, YAML, or CSV. Prior to 0.63.0, datamodel-code-generator validates a URL host once in src/datamodel_code_generator/http.py through get_body, _validate_url_for_fetch, and _get_ips_from_host, but then lets httpx resolve the host again for the connection, allowing DNS rebinding to bypass allow_private_network=False and reach internal services. This issue is fixed in version 0.63.0.

Official references

Related CVEs

CVECVSSEPSSKEVRExploitTitleMod.

CVE-2026-927197.5 HIG

—

———Quickwit through 0.9.0 fails to validate the host and scheme of the queue_url parameter in SQS file sources, allowing attackers to make the node issue requests to arbitrary internal addresses. Attackers can supply a malicious queue_url to the create-source API to scan internal networks and fingerprint services based on connection response differences.4h CVE-2026-926027.1 HIG

—

———TDuck survey form through version 5.3 fails to validate webhook URLs or verify form ownership in the WebhookConfigController. Authenticated attackers can attach webhooks to other users' forms and exfiltrate submissions to arbitrary external or internal addresses.5h CVE-2026-857324.7 MED

—

———oras-go is a Go library for managing OCI artifacts. Prior to 2.6.2, the parseLink function in registry/remote/utils.go accepts an absolute URL from a registry-controlled Link response header without validating its scheme, host, or port. Tags, Referrers, and Repositories pagination operations then issue a GET request to the attacker-selected URL from the victim's network, allowing blind server-side request forgery against internal services. The response body is not returned to the attacker, but timing and error differences can reveal service reachability, and credentials may be attached when the credential store has an entry for the target host. Exploitation requires a victim to perform a pagination-based listing operation against a malicious registry. The maintainer identifies this report as a duplicate of GHSA-3hr5-mjrr-hfjh and states that remediation is consolidated in that earlier advisory. The consolidated issue is fixed in version 2.6.2.4h CVE-2026-925694.3 MED

—

———Hippo4j through 1.5.0 contains a server-side request forgery vulnerability in four ThreadPoolController endpoints that fail to validate the clientAddress parameter. Authenticated attackers can supply arbitrary hostnames and ports to trigger outbound GET requests to internal networks and cloud metadata services.7h CVE-2026-925685.4 MED

—

———MLRun through 1.11.0 contains a server-side request forgery vulnerability in the WebhookNotification handler that allows authenticated users to make the API server send arbitrary HTTP requests to internal addresses. Attackers can update a run with a malicious webhook notification that executes when the run reaches a terminal state, enabling requests to internal services, Kubernetes APIs, or cloud metadata endpoints from within the cluster.6h CVE-2026-925668.2 HIG

—

———DataGear through 6.0.0 contains a server-side request forgery vulnerability in the /dataSet/preview/Http endpoint that allows unauthenticated attackers to execute arbitrary HTTP requests by supplying a caller-controlled URI. Attackers can issue GET, POST, PUT, PATCH, or DELETE requests to internal endpoints and cloud metadata services, receiving full response bodies without authentication or validation.6h