Base de CVEs

OpenClaw (npm package 'openclaw') versions >= 2026.4.5 and < 2026.8.1 can lose the originating requester's restrictions and untrusted provenance when session-derived text is persisted to session memory. In deployments where session-memory capture and dreaming are enabled, a restricted external sender whose messages are admitted with limited tools can persist instructions that are later supplied to an unattended background (dreaming) agent holding broader file and command capabilities, allowing actions beyond the authority of the original turn and affecting files, commands, or services available to that agent. Exploitation requires the content to be captured, selected for later processing, and followed by the model. The issue is fixed in 2026.8.1.

OpenClaw versions before 2026.8.1 fail to bind working directory context to reusable exec approvals, allowing approved commands to execute in different directories. Attackers with an allow-always approval can reuse it to run the same command against unreviewed files or repositories with materially different effects.

InvoicePlane is a self-hosted open source application for managing invoices, clients, and payments. Prior to 1.7.2, InvoicePlane fails to revoke administrative privileges after a role downgrade because Admin_Controller trusts the user_type snapshot stored in an existing session instead of revalidating ip_users.user_type. When one administrator downgrades another account, the target's active session continues to authorize administrative requests. The downgraded user can use Users::form() to set user_type back to 1, restoring the database role and making the privilege escalation persistent. This vulnerability is fixed in 1.7.2.

BerriAI LiteLLM before 1.101.0-rc.1 contains a tenant isolation bypass vulnerability in the semantic cache layer that allows authenticated users to read other tenants' cached responses by exploiting a metadata key mismatch between _get_semantic_cache_tenant_scope() and _get_metadata_variable_name(). Attackers holding a valid virtual key can submit semantically similar prompts on affected routes such as /v1/responses and /bedrock/* to retrieve cached responses containing other tenants' personally identifiable information, financial data, or source code, and can cause agentic front-ends to auto-execute attacker-supplied tool calls under victim credentials by returning cached function_call or tool_calls payloads to a different principal.

RabbitMQ is a messaging and streaming broker. From 3.13.0 until 3.13.18, 4.0.23, 4.1.14, 4.2.9, and 4.3.3, native MQTT and MQTT over WebSocket behind a trusted PROXY Protocol frontend could lose the proxy-derived client address before the MQTT authentication path checked loopback_users, causing the frontend-to-broker address to be treated as loopback. An attacker who can reach the trusted frontend and has valid credentials for a loopback-restricted account can therefore bypass the source-address restriction; the issue does not bypass password authentication. This issue is fixed in versions 3.13.18, 4.0.23, 4.1.14, 4.2.9, and 4.3.3.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.2, a customer who can view a ticket cannot see internal ticket articles through the article listing API. However, the same customer can directly request an attachment belonging to an internal article via the attachment download endpoint, bypassing article-level authorization. This results in an inconsistency: The article listing hides internal articles from customers. The attachment download only checks the parent ticket, not the article, so the same customer can download the attachment directly. This issue is fixed in version 7.0.2.

phpIPAM through 1.8.3 has incorrect authorization for id=="admins" and id=="all" in api/controllers/User.php.

The Botslab G980H dash camera firmware contains an authorization vulnerability in its session based command functionality. The product does not sufficiently associate an authenticated session with the client connection that established it, and subsequent privileged operations rely on possession of a valid session identifier without adequately validating the requesting client's authenticated context. An unauthenticated attacker with adjacent network access could potentially use valid session state associated with another client to access privileged functionality.

Incorrect Authorization vulnerability in TÜBİTAK ULAKBİM UlakPDF allows Authentication Bypass. This issue affects UlakPDF: through 09092026.

A missing authorization vulnerability exists in the Task Group APIs of Apache DolphinScheduler. The affected APIs do not properly verify whether the authenticated user has permission to access the project associated with the target Task Group. This issue affects Apache DolphinScheduler: before 3.4.3. Users are recommended to upgrade to version 3.4.3, which fixes the issue.

Klever-Go is the Go implementation of the Klever blockchain protocol. Prior to 1.7.20, the KleverUpdateAccountPermission built-in authorizes replacement of a target account's permissions by checking attacker-controlled vmInput.RecipientAddr instead of authenticated vmInput.CallerAddr. An attacker-controlled contract can choose a victim account with configured permissions as RecipientAddr, and contractHasValidPermission can accept the victim's default self-signer as authorization. UpdatePermission can then replace the victim's entire permission set with attacker-supplied Owner permissions, enabling asset theft or permanent lockout without a victim key or signature. Accounts without stored permissions and the native transaction path are not affected. This issue is fixed in version 1.7.20.

A flaw was found in sudo. When sudoers rules use NOTBEFORE or NOTAFTER time-based access restrictions with timestamps that omit the trailing 'Z' timezone indicator, the time evaluation relies on the TZ environment variable inherited from the calling user. Because sudo is a setuid-root program, an unprivileged local user can set TZ to an extreme timezone offset to shift the authorization window by up to approximately 25 hours, causing expired rules to be treated as valid. This allows the user to execute commands outside the intended time window. Authentication is not bypassed; only the time-based authorization check is affected.