Recipes CVE-2026-35046

This vulnerability presents moderate real-world risk despite a moderate CVSS score of 5.4. The CVSS vector (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:N) correctly reflects that exploitation requires network access, low complexity, authenticated user privilege, and user interaction on the client side, with cross-site scope impact. The UI:R (User Interaction Required) constraint significantly limits exploitability in automated attack scenarios. However, the EPSS score and exploitation probability should be weighted against the attack flow: an authenticated user with recipe edit permissions must deliberately inject malicious CSS, and the payload only activates when a client application renders the instructions_markdown field. This limits the threat to environments where (1) the Tandoor API is exposed to untrusted users with recipe editing capability, or (2) recipe data is consumed by third-party applications. The confirmed availability of a vendor patch (version 2.6.4) and the fix being already released further reduces active risk for patched deployments. Organizations using Tandoor Recipes for internal, trusted-user environments face minimal risk; those exposing the API to untrusted parties or relying on downstream consumers should prioritize patching. No active exploitation evidence is present in the available data.

An authenticated user with recipe editing permissions injects a malicious <style> tag containing CSS that redefines the login form position and opacity, overlaying a phishing form on top when another user views the recipe instructions via a web client. Alternatively, an attacker crafts CSS attribute selectors (e.g., input[value^="admin"]) to exfiltrate text via background-image URL calls to an attacker-controlled server. Since the attack requires authentication and client-side rendering of the instructions_markdown field, the threat is confined to organizations where the API is accessible to untrusted authenticated users or where recipe data feeds into third-party applications that do not implement additional sanitization.