CVE-2026-53708
MediumCVSS 6.6Exploitation Probability (EPSS)
Low risk26th percentile - higher than 26% of all known CVEs
Summary
ContextForge is an AI gateway, registry, and proxy that provides centralized discovery, guardrails, and management for MCP, A2A, and REST or gRPC APIs. Prior to 1.0.3, the /admin/gateways/test call site in mcpgateway/admin.py calls validate_gateway_test_url() in mcpgateway/common/validators.py to resolve and reject private, loopback, link-local, and cloud-metadata addresses, but ResilientHttpClient later resolves the original hostname again without binding the validated address. When MCPGATEWAY_ADMIN_API_ENABLED is enabled, an attacker with a database-backed role containing explicit gateways.read permission can use DNS rebinding to return a public address during validation and a private or metadata address during connection, bypassing ssrf_blocked_networks and ssrf_dns_fail_closed because those controls apply only to the validation-time result. The endpoint's allow_admin_bypass=False setting means a bootstrap-only virtual platform-admin identity without a database role is not sufficient. Successful exploitation can reach internal services and cloud metadata, expose cloud credentials, access internal APIs, or probe internal network ports. This issue is fixed in version 1.0.3.
Risk Assessment
An attacker can bypass SSRF protections and access internal resources, potentially leading to cloud credential exposure or system integrity compromise.
Recommendation
Update ContextForge to version 1.0.3 or later.
Related vulnerabilities
- CVE-2026-77596Critical
In Kubeflow Pipelines prior to 2.18.0, the multi-user HTTP endpoints for uploading pipelines and pipeline versions skip SubjectAccessReview when the target namespace is empty and REQUIRE_NAMESPACE_FOR_PIPELINES defaults to false. A low-privilege tenant can create shared pipelines or inject attacker-controlled versions into existing shared pipelines without the required write permission. If the API server is directly reachable without an authenticating proxy, an unauthenticated caller can exploit the same bypass.
- CVE-2026-119757Critical
BTCPay Server before 2.4.2, including 2.4.2 release candidates, contains an information disclosure vulnerability that allows unauthenticated remote attackers to retrieve .macaroon credential files for an LND node. The exposed macaroon can be used to gain full control of the LND node, including moving funds, resulting in theft of funds.
- CVE-2026-111494Critical
The MatchPoint Android app incorrectly validates TLS certificates due to a TrustManager that accepts any certificate and a permissive hostname validator. The app also allows unencrypted HTTP traffic, enabling an attacker who intercepts traffic to read and modify communications, including credentials and personal data.
- CVE-2026-111490Critical
An insecure direct object reference (IDOR) vulnerability in the MatchPoint API relies on the `actingasuser` HTTP header without verifying the requester's authorization for the specified user. Manipulation of sequential identifiers allows access to personal, banking, and financial data, private messages, and physical-access QR codes, as well as modification of third-party accounts, password changes, impersonation, and virtual balance transfers.
- CVE-2026-111489Critical
A static credential embedded in the MatchPoint mobile app has been reported to return the 'autorizado' token. The report states that this token grants unauthenticated access to API services.
- CVE-2026-85076Critical
Prior to 0.9.124, an authenticated non-admin with run_events, create_events, or edit_events privilege in Cronicle can supply undeclared keys in event params through api_create_event, api_update_event, or api_run_event. A key colliding with a security-sensitive environment variable can override the plugin execution context, potentially causing privilege escalation, arbitrary code execution, data access, or service disruption.
- CVE-2026-48705Critical
Prior to 1.6.1, MONAI's PydicomReader._get_affine() trusts non-finite values in DICOM PixelSpacing, ImagePositionPatient, and ImageOrientationPatient fields and composes them into an affine matrix without validation. An unauthenticated attacker who can upload a crafted DICOM to a default MONAILabel deployment can propagate NaN or infinity values, causing inference failures, persistent datastore poisoning, or geometrically corrupted results.
- CVE-2026-46350Critical
Prior to 0.52.0, dcmjs's DecimalString and IntegerString implementations in src/ValueRepresentation.js sanitize text with a permissive regular expression and pass the result to JavaScript's Number constructor without validating finiteness or DICOM rules. Scientific notation overflow in a DS or IS field can produce Infinity or NaN, and IntegerString accepts exponent notation and values outside the signed 32-bit IS range. Applications trusting these values can propagate invalid numbers into geometry, windowing, dose scaling, and volume reconstruction.
- CVE-2026-116790Critical
Prior to 1.11.2, FreeSWITCH's inbound ICE handling accepts every connectivity check without cryptographically validating the STUN MESSAGE-INTEGRITY attribute. A network peer can forge an ICE connectivity check without SDP credentials, nominate its own address, redirect RTP and DTLS traffic, and terminate the legitimate media path, enabling eavesdropping and media injection on unencrypted legs.
- CVE-2026-116785Critical
Prior to 1.11.2, FreeSWITCH's ws_read_frame() decodes a WebSocket frame's 64-bit extended payload length with a 32-bit byte swap and narrows the result into a signed length. A value whose surviving low word becomes negative bypasses the signed size check and is reinterpreted as an enormous unsigned count, causing client bytes to overwrite memory beyond the 64 KiB receive buffer. This can corrupt connection pointers and crash FreeSWITCH, ending all concurrent sessions.
Original NVD description (English source)
ContextForge is an AI gateway, registry, and proxy that provides centralized discovery, guardrails, and management for MCP, A2A, and REST or gRPC APIs. Prior to 1.0.3, the /admin/gateways/test call site in mcpgateway/admin.py calls validate_gateway_test_url() in mcpgateway/common/validators.py to resolve and reject private, loopback, link-local, and cloud-metadata addresses, but ResilientHttpClient later resolves the original hostname again without binding the validated address. When MCPGATEWAY_ADMIN_API_ENABLED is enabled, an attacker with a database-backed role containing explicit gateways.read permission can use DNS rebinding to return a public address during validation and a private or metadata address during connection, bypassing ssrf_blocked_networks and ssrf_dns_fail_closed because those controls apply only to the validation-time result. The endpoint's allow_admin_bypass=False setting means a bootstrap-only virtual platform-admin identity without a database role is not sufficient. Successful exploitation can reach internal services and cloud metadata, expose cloud credentials, access internal APIs, or probe internal network ports. This issue is fixed in version 1.0.3.
Vulnerability data from NVD (NIST) · CISA KEV · EPSS

