CVE-2026-27825
Published: 10 March 2026
Summary
CVE-2026-27825 is a critical-severity Path Traversal (CWE-22) vulnerability in Mcp-Atlassian Mcp Atlassian. Its CVSS base score is 9.0 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Cron (T1053.003); ranked at the 5.9th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
This vulnerability is AI-related — categorised as AI Agent Protocols and Integrations; in the Protocol-Specific Risks risk domain.
The strongest mitigations our analysis identified are NIST 800-53 AC-3 (Access Enforcement) and SI-10 (Information Input Validation).
Threat & Defense at a Glance
Threat & Defense Details
Mitigating Controls (NIST 800-53 r5)AI
SI-10 requires validation of the download_path parameter to enforce directory boundaries and block path traversal attempts.
AC-3 mandates enforcement of access authorizations that restrict file writes to approved directories, directly countering the lack of boundary enforcement in the MCP tool.
AC-6 limits server process privileges to the minimum necessary, reducing the impact of arbitrary file writes by preventing access to sensitive paths like /etc/cron.d/.
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
Path traversal enables arbitrary file write to server-writable locations (e.g., /etc/cron.d/), directly facilitating cron-based scheduled execution for RCE; exploitation of the vuln with low privileges maps to privilege escalation.
NVD Description
MCP Atlassian is a Model Context Protocol (MCP) server for Atlassian products (Confluence and Jira). Prior to version 0.17.0, the `confluence_download_attachment` MCP tool accepts a `download_path` parameter that is written to without any directory boundary enforcement. An attacker who can…
more
call this tool and supply or access a Confluence attachment with malicious content can write arbitrary content to any path the server process has write access to. Because the attacker controls both the write destination and the written content (via an uploaded Confluence attachment), this constitutes for arbitrary code execution (for example, writing a valid cron entry to `/etc/cron.d/` achieves code execution within one scheduler cycle with no server restart required). Version 0.17.0 fixes the issue.
Deeper analysisAI
CVE-2026-27825 is a path traversal vulnerability (CWE-22, CWE-73) in the MCP Atlassian server, a Model Context Protocol (MCP) implementation for Atlassian products including Confluence and Jira. In versions prior to 0.17.0, the `confluence_download_attachment` MCP tool accepts a `download_path` parameter without directory boundary enforcement, allowing arbitrary file writes to paths accessible by the server process. This flaw enables attackers to overwrite files with content sourced from a Confluence attachment, potentially leading to arbitrary code execution. The vulnerability carries a CVSS v3.1 base score of 9.0 (AV:A/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H).
An attacker with low privileges (PR:L) and adjacent network access (AV:A) can exploit this by calling the vulnerable MCP tool and providing a malicious `download_path` along with a Confluence attachment containing attacker-controlled content. Successful exploitation allows writing arbitrary data to any server-writable location, such as creating a cron job in `/etc/cron.d/` to achieve remote code execution within one scheduler cycle without requiring a server restart. No user interaction is needed (UI:N), and the scoped impact (S:C) amplifies confidentiality, integrity, and availability effects.
The issue is fixed in MCP Atlassian version 0.17.0. Security practitioners should upgrade to this version immediately. Relevant advisories and patch details are available in the GitHub security advisory at https://github.com/sooperset/mcp-atlassian/security/advisories/GHSA-xjgw-4wvw-rgm4 and the fixing commit at https://github.com/sooperset/mcp-atlassian/commit/52b9b0997681e87244b20d58034deae89c91631e.
Details
- CWE(s)
Affected Products
AI Security AnalysisAI
- AI Category
- AI Agent Protocols and Integrations
- Risk Domain
- Protocol-Specific Risks
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- Matched keywords: mcp, model context protocol, mcp, mcp