Typebot CVE-2026-28444

The attacker must be authenticated to the Typebot platform with at least a low-privilege account (CVSS PR:L confirmed). The vulnerability is triggered by calling the getResultLogs API endpoint with any valid resultId from any workspace; the attacker needs their own typebotId (from any bot they control) to pass the authorization check. The victim's resultId must be known or enumerable - Typebot result IDs are typically UUIDs, so bulk enumeration is impractical but leakage via webhooks, browser history, or other API responses is plausible. No special server configuration or non-default feature flag is required - the vulnerable code path is active in all default Typebot deployments running an affected version. Exploitation is limited to confidentiality impact only (read-only); no write or delete operations are exposed through this flaw.

CVSS 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N) correctly reflects a network-reachable, low-complexity, authenticated attack with high confidentiality impact and no integrity or availability effect. The PR:L vector is the primary limiting factor - the attacker must already hold a valid Typebot account, which restricts the exposed population to registered or compromised platform users. Attack complexity is genuinely low once a resultId is obtained; UUIDs used as primary keys may be guessable or leakable through other API interactions (e.g., webhook logs, URL parameters). No EPSS data was provided; given the absence of KEV listing and no identified POC, opportunistic mass exploitation appears unlikely at this time, but targeted abuse by malicious tenants on shared Typebot instances is a realistic concern. The SSVC technical impact is rated partial (confidentiality only, scoped to one result per request), but webhook payloads and AI outputs can contain PII, credentials, or business-sensitive data, elevating real-world severity above the raw CVSS score suggests. The oversight nature of the bug - confirmed by the fact that all peer endpoints already perform the correct check - means it is likely straightforward to exploit without specialized knowledge.

An attacker with a legitimate low-privilege Typebot account sends a GET request to the getResultLogs endpoint, substituting their own valid typebotId for authorization while supplying a victim's resultId (obtained via enumeration, intercepted webhook, or leaked URL). The server authorizes the request against the attacker's typebot, fetches the log record by resultId alone, and returns the full execution log - potentially including webhook response bodies, AI model outputs with customer data, and internal integration payloads - without any cross-ownership check. No public proof-of-concept code has been identified, but the attack requires only standard HTTP tooling and knowledge of the API contract.