Nlgbao1340

Dokploy 0.28.0 before 0.31.1 contains a path traversal vulnerability that allows authenticated organization members to read arbitrary files by supplying ../ sequences in the filePath of patch.readRepoFile. Attackers with service read permission can read /etc/shadow, SSH private keys, and secret environment files on the Dokploy host or managed remote servers.

The Local Gateway desktop app contains a flaw in its deep-link connect handler that fails to require user confirmation. This allows the application to be silently re-paired with a different n8n instance. An attacker can use a crafted `n8n-computer-use://` link to connect the user's computer to an instance under the attacker's control, allowing the attacker to inherit previously granted permissions and execute actions on the computer via the gateway.

An authorization bypass exists in the agent chat resume endpoint. The system accepts a run identifier from the request body without verifying if the suspended checkpoint belongs to the caller. This allows authenticated project members to list and read private agent conversations of other members via project-scoped endpoints and approve pending tool calls, leading to tool execution and the writing of results into the victim's conversation.

A credentials leak exists in the Wekan and Baserow username-and-password credentials. This issue allows the transmission of unencrypted passwords to unvalidated hosts. An attacker with permissions to update credentials can modify the `host` field to redirect account passwords to arbitrary hosts, effectively bypassing domain validation controls.

A path traversal issue exists in the n8n node due to improper validation of resource identifiers. This allows attackers to use malicious resource IDs to redirect API calls to unintended resources, potentially granting unauthorized access to workflows, executions, and credential secrets within the scope of the API key.

Session token leakage occurs in the Dynamic Credentials 'authorize' and 'revoke' endpoints. Attackers with resolver registration capability can capture session tokens of collaborators by configuring a fallback resolver to point to an attacker-controlled endpoint during the account connection flow, which allows unauthorized access to credentials.

Insufficient per-project authorization in the Insights API routes allows authenticated users with insights scopes to access workflow names and execution statistics across different projects. An attacker can provide arbitrary `projectId` parameters to retrieve sensitive project and workflow information from projects to which they do not belong.

An authenticated attacker with Editor access or `alert.instances.external:write` permissions can submit an external Alertmanager alert containing a controlled `generatorURL`. The software renders `alert.generatorURL` directly as the href of the See source LinkButton without proper URL-scheme sanitization or a safe-protocol allowlist. The click interceptor's :// heuristic can be bypassed by placing :// inside a JavaScript comment. Consequently, when a user with read access clicks the See source link, the browser executes attacker-controlled JavaScript within the Grafana origin using the permissions of the clicking user.

Multiple AI and LLM nodes fail to enforce the Allowed HTTP Request Domains allowlist when base or endpoint URLs are configured by the user. This allows low-privileged workflow editors who have use-only access to shared credentials to redirect requests to hosts controlled by an attacker, leading to the exfiltration of credential secrets for subsequent use against underlying services.

Insufficient validation of credential access occurs when using the Execute Sub-workflow node. The system only validates top-level credentials and fails to check credentials referenced within the inline workflow JSON. A user with Editor access to a shared workflow can reference a target credential ID within the inline JSON to bypass save-time and runtime validation. This allows the user to utilize or exfiltrate credentials from the parent workflow's project context that they are not authorized to access.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, the onebox method in the SharedAiConversation model renders the conversation title directly into HTML without proper sanitization. This could allow for potential issues related to the display of conversation titles.

Discourse, an open-source discussion platform, is affected by a cross-site scripting issue. The system improperly trusts and renders output from an AI Large Language Model (LLM) using `htmlSafe` within the Review Queue interface, lacking sufficient sanitization. An attacker can leverage Prompt Injection techniques to make the AI generate a malicious payload, such as tags. When a staff member (Admin/Moderator) views a flagged post in the Review Queue, this payload is executed.

Outline is a collaborative documentation service. A security issue exists in the document restoration logic that allows any team member to restore, view, and take ownership of deleted drafts belonging to other users, including administrators, without authorization. This is due to a bypass of ownership validation during the restoration process, potentially allowing an attacker to access sensitive private information and prevent the original owner from accessing their content. The issue does not affect versions 1.4.0 and later.

Outline is a collaborative documentation service. Prior to version 1.5.0, a logic flaw exists in the filtering mechanism of the `/events.list` API endpoint, which is used to retrieve activity logs. This flaw allows any authenticated user to access activity events related to documents lacking a collection (such as Private Drafts and Deleted Documents), bypassing standard permission controls. While the document content itself is not directly exposed, sensitive metadata is leaked, including Document IDs, user activity timestamps, and, in some cases, the Document Title of Permanent Delete. The leakage of valid Document IDs from deleted drafts compromises the randomness of UUIDs, simplifying the exploitation of high-severity IDOR attacks, like those affecting the `documents.restore` function.

Flowise, a drag & drop user interface for building customized large language model flows, contains a Server-Side Request Forgery (SSRF) issue. The application exposes an HTTP Node within AgentFlow and Chatflow that makes server-side HTTP requests using URLs controlled by the user. There are no restrictions on the target hosts, allowing requests to private IP ranges, localhost, or cloud metadata endpoints. This enables an attacker interacting with a publicly exposed chatflow to force the Flowise server to make requests to internal network resources that are otherwise inaccessible from the public internet.

n8n is a workflow automation platform. A flaw in the Merge node’s SQL Query mode permitted authenticated users with workflow creation or modification rights to write arbitrary files to the n8n server’s filesystem, potentially resulting in remote code execution. The issue affects users with the ability to create or modify workflows.

n8n is a workflow automation platform. A stored Cross-Site Scripting (XSS) issue may occur when using the “Respond to Webhook” node in versions before 1.114.0. If this node responds with HTML content containing executable scripts, the payload may execute directly in the top-level window. This can allow a malicious actor with workflow creation permissions to execute arbitrary JavaScript in the n8n editor interface.