These pages are published after PatchSiren validates generated defensive summaries against stored public CVE and source evidence.
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows for markup or script injection in the browser of a user completing the OAuth flow due to improper escaping of the OAuth error query parameter. This issue is fixed in version 0.22.0. The vulnerability impacts users completing the OAuth flow, and defenders should assess exposure and apply the patch to prev [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to upload arbitrary local files by providing an absolute or traversal file path. This issue, fixed in version 0.22.0, was initially addressed as part of the remediation for CVE-2026-27825 but was not fully constrained. The vulnerability affects the upload_attachment, file_path, and CVE-2026-2 [truncated]
CVE-2026-77268 MCP Atlassian OAuth fallback token directory and JSON file created without explicit owner-only modes, allowing local users or processes to read access and refresh tokens and reuse the associated Atlassian session. System administrators and security teams should review local access controls and token management practices, and update to version 0.22.0 or later. This issue affects MCP Atlassia [truncated]
CVE-2026-77266 is a vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, which allows an MCP caller with attachment access to read a chosen server-local file and exfiltrate it through Jira or Confluence. The issue is fixed in version 0.22.0. Defenders should assess exposure and apply the patch to prevent server-local file exfiltration. The vulnerability involves pa [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to upload arbitrary server-readable files to Confluence by exploiting the confluence_upload_attachment function with a controlled file_path. This issue, fixed in version 0.22.0, enables attackers to traverse outside the workspace. Defenders should assess exposure, particularly those managing [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to upload and retrieve server-local files without proper authorization. This issue, fixed in version 0.22.0, enables the upload of environment files, credentials, or other readable host data to a Confluence page. The vulnerability is caused by the confluence_upload_attachment function not ens [truncated]
A remote MCP caller with tool access can cause the server to read sensitive local files and upload them as Atlassian attachments due to a vulnerability in MCP Atlassian prior to 0.22.0. The issue is fixed in version 0.22.0. This vulnerability allows attackers to access sensitive information and potentially use it for malicious purposes. Defenders should assess exposure and apply the patch to prevent unaut [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows local users to read a plaintext OAuth fallback file containing refresh and access tokens due to inherited permissions from the process umask. This issue is fixed in version 0.22.0. The vulnerability affects local users who can read the backup and retain Atlassian access through the refresh token. Defende [truncated]
The CVE-2026-77255 vulnerability in MCP Atlassian allows an attacker to read arbitrary local files and attach them to a Jira issue, potentially exfiltrating sensitive information. This issue is fixed in version 0.22.0. The vulnerability arises from the Jira update_issue attachments argument being converted into local paths and routed to the attachment upload implementation without workspace validation. A [truncated]
A critical vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows unauthenticated network callers to perform operations with the operator account's permissions. This issue, fixed in version 0.22.0, enables attackers to bypass authentication and leverage globally configured Jira or Confluence credentials.
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows disclosure of configuration, credentials, and other sensitive files readable by the MCP process. The issue, fixed in version 0.22.0, arises from the acceptance of arbitrary local filesystem paths in Jira and Confluence attachment upload tools. This vulnerability can lead to potential disclosure of sensit [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to bypass redirect checks and potentially access internal addresses. The issue, fixed in version 0.22.0, arises from the use of the module-level requests.get function instead of the fetcher's protected session in JiraUserMixin._lookup_user_by_permissions.
An unauthenticated network caller can read files available to the MCP process, upload them to an attacker-selected Jira issue or Confluence page, and retrieve the contents due to vulnerabilities in MCP Atlassian versions prior to 0.22.0. This issue allows for potential data exfiltration or unauthorized file uploads, compromising the integrity of the MCP process. Defenders should assess exposure and priori [truncated]
CVE-2026-77247 is a vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, which allows permitted clients to disclose host files without shell or direct filesystem access. The issue arises from the interpretation of caller-controlled path arguments on the MCP server, which opens files before sending them as attachments. This vulnerability is fixed in version 0.22.0. [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to send a server-local file to a selected attachment endpoint without proper authorization. This issue is fixed in version 0.22.0. The vulnerability arises from an HTTP transport deployment with READ_ONLY_MODE=false that accepts a request without an Authorization identity and permits attacker [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows server-side requests to protected network resources due to a backslash authority confusion in the validate_url_for_ssrf function. This issue is fixed in version 0.22.0. The vulnerability can lead to unauthorized access to internal network resources, potentially resulting in data breaches or system compro [truncated]
CVE-2026-77271 MCP Atlassian module overwrite allows code execution prior to version 0.22.0. This vulnerability affects Confluence and Jira deployments using MCP Atlassian, allowing for code execution via module overwrite. Defenders should assess exposure and prioritize remediation due to the potential for significant consequences if exploited, including attacker-selected writes within the working directo [truncated]
CVE-2026-77270 MCP Atlassian server vulnerability allows file disclosure via Jira and Confluence attachment upload tools treating caller-controlled file paths as trusted server-local paths, enabling an attacker to disclose any file readable by the server process. This issue is fixed in version 0.22.0. Defenders should assess exposure, prioritize remediation, and monitor for suspicious activity related to [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to bypass incomplete CVE-2026-27826 remediation by supplying internal or metadata-service URLs via specific headers, potentially leading to SSRF attacks. The issue is fixed in version 0.22.0. Defenders should assess exposure and apply the patch to prevent potential SSRF attacks. Evidence is l [truncated]
An unauthenticated caller can use a DNS-rebinding hostname that returns a public address during validation and an internal address during connection, causing requests to internal or metadata services in MCP Atlassian prior to 0.22.0. This issue allows for potential unauthorized access and requires defenders to assess exposure and prioritize upgrading to version 0.22.0 or later. The vulnerability is trigge [truncated]
MCP Atlassian is a Model Context Protocol (MCP) server for Atlassian products (Confluence and Jira). CVE-2026-77261 is a Server-Side Request Forgery (SSRF) vulnerability in MCP Atlassian instances prior to version 0.22.0. The vulnerability allows an attacker-controlled or compromised configured Atlassian instance to return a redirect to an internal address, potentially leading to SSRF attacks. Defenders s [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows a permitted MCP caller to upload sensitive host files to an Atlassian destination and retrieve their contents. The issue, fixed in version 0.22.0, is caused by the Confluence and Jira upload_attachment implementations accepting an unconstrained file_path and opening the referenced server-local file.
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows a permitted Confluence MCP caller to upload a file as an attachment and disclose data readable by the server process. The issue is fixed in version 0.22.0. This vulnerability impacts data confidentiality and requires defenders to assess exposure and apply the patch to prevent potential data disclosure. T [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows an attacker to bypass administrator-configured allowlists and search projects or spaces outside the intended boundary. This issue is fixed in version 0.22.0. The vulnerability can be exploited by manipulating the projects_filter and spaces_filter arguments, potentially leading to unauthorized access. Def [truncated]
The CVE-2026-77251 issue involves MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products like Confluence and Jira. Prior to version 0.22.0, the server has vulnerabilities in Jira search, Confluence search, and Jira board APIs. These vulnerabilities allow exposure of issues, boards, or pages outside JIRA_PROJECTS_FILTER or CONFLUENCE_SPACES_FILTER when operator credentials have broader [truncated]
CVE-2026-77250 MCP Atlassian OAuthConfig Token Exposure. The vulnerability exists in MCP Atlassian's OAuthConfig, which writes plaintext fallback files containing access and refresh tokens under the user's .mcp-atlassian directory with process-default permissions. On systems with a permissive umask, local users and processes can read these tokens, potentially leading to unauthorized Atlassian access. Syst [truncated]
A critical vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows unauthenticated access to Atlassian tools as the operator. The issue, fixed in version 0.22.0, enables a network client to invoke read and write operations available to the operator's account. This vulnerability is particularly concerning because it allows for unauthorized access to sensitive da [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows unauthorized tool invocation. This issue, fixed in version 0.22.0, enables a client with knowledge of a hidden tool name to bypass configured least-privilege restrictions and directly invoke excluded read, write, or delete tools. The vulnerability arises from ENABLED_TOOLS and TOOLSETS being applied when [truncated]
A vulnerability in MCP Atlassian, a Model Context Protocol (MCP) server for Atlassian products, allows unauthenticated access to internal or metadata endpoints. The issue arises from the validate_url_for_ssrf function's validation of a hostname's resolved addresses, which can be bypassed using a short-lived DNS answer that is public during validation and private during connection. This vulnerability was f [truncated]
CVE-2026-73498 is a high-severity vulnerability in the MCP Atlassian server, allowing authenticated clients to read and exfiltrate server files to Confluence as attachments. The issue arises from the confluence_upload_attachment function passing client-supplied file paths directly to open() without proper validation. This flaw can also be exploited through AI agents if they can be induced to call the tool [truncated]
