CVE WatchOct 6, 2026

CVE-2026-103923

KaTeX is a fast, easy-to-use JavaScript library for TeX math rendering on the web. From 0.11.0 until 0.18.2, KaTeX uses ordinary JavaScript

CVSS

—

No CVSS

EPSS

0.3%

p22

KEV

—

Exploit Today

0-100

Published: Oct 1, 2026 · Last modified: Oct 6, 2026 · CWE-807

EPSS · 30d

0.3%EPSS · 30 days0.3%

2026-10-022026-10-25

Technical description

KaTeX is a fast, easy-to-use JavaScript library for TeX math rendering on the web. From 0.11.0 until 0.18.2, KaTeX uses ordinary JavaScript property access for the renderer options object, the trust setting, default and processor setting metadata, and namespace lookup and group restoration, allowing inherited properties to be treated as explicitly supplied values. When Object.prototype is already polluted or an attacker controls the options object's prototype, attacker-controlled mathematical expressions can use an inherited trust value to enable trusted rendering and produce links capable of user-interaction cross-site scripting or loading attacker-selected external resources in a consuming application that inserts unsanitized KaTeX output into a page. KaTeX does not itself create the prototype pollution, and rendering an expression alone does not execute script. This issue is fixed in version 0.18.2.

Official references

Related CVEs

CVECVSSEPSSKEVRExploitTitleMod.

CVE-2026-1160167.1 HIG

16.9%

——5yudao-cloud through 2026.09 contains an authorization bypass vulnerability that allows authenticated users to evade row-level data scope restrictions by sending a data-permission-enable header set to false. Attackers can add this header to /admin-api requests, which DataPermissionRpcWebFilter trusts, to read or modify other departments' and users' records within the same tenant.1d CVE-2026-1146174.2 MED

14.0%

——4Authlib is a Python library which builds OAuth and OpenID Connect servers. Prior to 1.9.0, OpenIDMixin.parse_id_token() reads the nonce_supported decision from the untrusted ID token being validated and sets the expected nonce to null when that claim is the JSON boolean false. In implicit and hybrid relying parties, or applications that directly accept externally supplied ID tokens, an attacker with a captured validly signed token can bypass the nonce-to-session binding and replay the victim's token into an attacker-controlled session. Plain authorization code flow is not affected, and the token must still pass signature, issuer, and audience validation. OpenIDMixin.parse_id_token() trusts nonce_supported in both authlib/integrations/base_client/sync_openid.py and authlib/integrations/base_client/async_openid.py to disable the expected nonce. This issue is fixed in version 1.9.0.2d CVE-2026-501066.5 MED

10.7%

——3When NLnet Labs Unbound up to and including version 1.26.1 is configured with both a root trust anchor and an additional 'trust-anchor:' for an island-of-security zone whose parent is provably insecure under the root chain, a malicious actor who can deliver one forged NXDOMAIN response for a name inside the island, with the NSEC/SOA RRSIG 'SignerName' set to the insecure parent, tricks Unbound's validator to use the root trust anchor instead, and consecutively the insecure parent, thus causing that record to be treated as insecure and denying existence for records in the trust island.3d CVE-2026-111748—

32.4%

——10AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to 3.14.4, an attacker can send an absolute-form request that causes Request.scheme to use attacker-controlled request-target scheme data instead of the connection protocol. Applications that use Request.scheme for security decisions can consequently be tricked into treating an unencrypted connection as encrypted or an encrypted connection as unencrypted. This issue is fixed in version 3.14.4.6d CVE-2026-1118695.3 MED

44.3%

——13A vulnerability was identified in makeplane plane up to 1.4.2. Impacted is an unknown function of the component API Endpoint. Such manipulation of the argument X-Forwarded-For leads to reliance on untrusted inputs in a security decision. It is possible to launch the attack remotely. The exploit is publicly available and might be used.4d CVE-2026-1084337.8 HIG

35.9%

——11Reliance on untrusted inputs in a security decision in Microsoft Office Word allows an unauthorized attacker to bypass a security feature locally.5d