CVE-2026-77267: MCP Atlassian: SSRF via header URLs bypasses prior fix

Q: Is CVE-2026-77267 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2026-77267, increasing the risk of exploitation.

Q: How to fix CVE-2026-77267?

1) Upgrade mcp-atlassian to 0.22.0 or later (the fix is in PR #1448). 2) Until patched, do not expose the server on untrusted networks; bind it to localhost or place it behind an authenticated reverse proxy that strips or overwrites the X-Atlassian-Jira-Url and X-Atlassian-Confluence-Url headers. 3) Apply egress filtering to block the server's access to RFC1918 ranges, link-local addresses and metadata services. 4) Enforce IMDSv2 or disable the metadata service where possible. 5) Detection: search access and proxy logs for these headers carrying hosts other than your Atlassian tenant, and for outbound connections from the server to internal or link-local addresses. 6) Inventory all MCP servers and agent tool hosts running mcp-atlassian, including developer workstations and containers.

Q: What systems are affected by CVE-2026-77267?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, MCP servers, tool-calling LLM integrations, enterprise Jira/Confluence integrations.

Q: What is the CVSS score for CVE-2026-77267?

No CVSS score has been assigned yet.

GHSA-5wf4-jqxh-8gm3HIGHPoC AVAILABLECISA: TRACK*

Published September 22, 2026

CISO Take

MCP Atlassian, the MCP server that connects AI agents to Jira and Confluence, fails to validate the X-Atlassian-Jira-Url and X-Atlassian-Confluence-Url headers, so any caller who can set them can make the server send requests to internal hosts or cloud metadata services. This bypasses the earlier CVE-2026-27826 fix, and the package already has 38 other CVEs, which points to weak security hygiene. Exploitation is plausible: a public PoC exists and SSVC says TRACK*, though EPSS is only 0.00294 (top 80%), the CVE is not in CISA KEV, and no Nuclei template exists. Downstream blast radius is small (1 dependent), but the SSRF can reach cloud credentials where the server runs in a cloud environment. Upgrade to mcp-atlassian 0.22.0 or later, and restrict network access to the server and its egress to metadata endpoints until you do.

Sources:NVDEPSSGitHub AdvisoryATLASCISA SSVC

What is the risk?

Rated high. The flaw is a network-reachable SSRF (CWE-918) with a public PoC and a trivial trigger (setting an HTTP header). The CVSS vector is not published and the score is unavailable. Exploitation likelihood is moderate to low by EPSS (0.00294), and the CVE is not in KEV. Risk rises sharply when the server runs in HTTP/SSE transport mode, is reachable beyond localhost, or sits on a cloud host with an accessible metadata service (for example 169.254.169.254). Stdio-only local deployments are far less exposed. A previous remediation was incomplete, which lowers confidence in the maintainers' validation coverage.

How does the attack unfold?

Initial Access

Attacker reaches the exposed mcp-atlassian HTTP endpoint, or influences headers sent by an agent client.

AML.T0049

Exploitation

A crafted X-Atlassian-Jira-Url or X-Atlassian-Confluence-Url header pointing to an internal or metadata URL is accepted without SSRF validation.

AML.T0049

Credential Access

The server fetches the internal resource and returns cloud credentials or internal data to the attacker.

AML.T0055

Impact

Stolen credentials or internal access enable lateral movement into cloud and internal systems.

Initial Access

Attacker reaches the exposed mcp-atlassian HTTP endpoint, or influences headers sent by an agent client.

AML.T0049

Exploitation

A crafted X-Atlassian-Jira-Url or X-Atlassian-Confluence-Url header pointing to an internal or metadata URL is accepted without SSRF validation.

AML.T0049

Credential Access

The server fetches the internal resource and returns cloud credentials or internal data to the attacker.

AML.T0055

Impact

Stolen credentials or internal access enable lateral movement into cloud and internal systems.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
MCP Atlassian	pip	< 0.22.0	`0.22.0`
2 dependents85% patched~4d to patchFull package profile →

Do you use MCP Atlassian? You're affected.

How severe is it?

CVSS 3.1

N/A

EPSS

0.3%

chance of exploitation in 30 days

Higher than 25% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○CISA SSVC: Public PoC

○Public PoC indexed (trickest/cve)

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What should I do?

1 step

1) Upgrade mcp-atlassian to 0.22.0 or later (the fix is in PR #1448). 2) Until patched, do not expose the server on untrusted networks; bind it to localhost or place it behind an authenticated reverse proxy that strips or overwrites the X-Atlassian-Jira-Url and X-Atlassian-Confluence-Url headers. 3) Apply egress filtering to block the server's access to RFC1918 ranges, link-local addresses and metadata services. 4) Enforce IMDSv2 or disable the metadata service where possible. 5) Detection: search access and proxy logs for these headers carrying hosts other than your Atlassian tenant, and for outbound connections from the server to internal or link-local addresses. 6) Inventory all MCP servers and agent tool hosts running mcp-atlassian, including developer workstations and containers.

What does CISA's SSVC say?

DecisionTrack*

Exploitationpoc

AutomatableNo

Technical Impactpartial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Data ExtractionAuth BypassAgentPluginAPIAML.T0049- Exploit Public-Facing Application AML.T0055- Unsecured Credentials AML.T0106- Exploitation for Credential Access

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art. 15-Accuracy, robustness and cybersecurity

ISO 42001

A.10.3-Suppliers

NIST AI RMF

MANAGE 2.3-Respond to and recover from previously unknown risks

OWASP LLM Top 10

LLM03-Supply ChainLLM06-Excessive Agency

Frequently Asked Questions

What is CVE-2026-77267?

Is CVE-2026-77267 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2026-77267, increasing the risk of exploitation.

How to fix CVE-2026-77267?

1) Upgrade mcp-atlassian to 0.22.0 or later (the fix is in PR #1448). 2) Until patched, do not expose the server on untrusted networks; bind it to localhost or place it behind an authenticated reverse proxy that strips or overwrites the X-Atlassian-Jira-Url and X-Atlassian-Confluence-Url headers. 3) Apply egress filtering to block the server's access to RFC1918 ranges, link-local addresses and metadata services. 4) Enforce IMDSv2 or disable the metadata service where possible. 5) Detection: search access and proxy logs for these headers carrying hosts other than your Atlassian tenant, and for outbound connections from the server to internal or link-local addresses. 6) Inventory all MCP servers and agent tool hosts running mcp-atlassian, including developer workstations and containers.

What systems are affected by CVE-2026-77267?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, MCP servers, tool-calling LLM integrations, enterprise Jira/Confluence integrations.

What is the CVSS score for CVE-2026-77267?

No CVSS score has been assigned yet.

What is the AI security impact?

Affected AI Architectures

agent frameworksMCP serverstool-calling LLM integrationsenterprise Jira/Confluence integrations

MITRE ATLAS Techniques

AML.T0049Exploit Public-Facing Application

AML.T0055Unsecured Credentials

AML.T0106Exploitation for Credential Access

Compliance Controls Affected

EU AI Act: Art. 15

ISO 42001: A.10.3

NIST AI RMF: MANAGE 2.3

OWASP LLM Top 10: LLM03, LLM06

What are the technical details?

Original Advisory

MCP Atlassian is a Model Context Protocol (MCP) server for Atlassian products (Confluence and Jira). Prior to 0.22.0, the X-Atlassian-Jira-Url and X-Atlassian-Confluence-Url headers are processed by _process_authentication_headers and used to construct Atlassian fetchers without calling validate_url_for_ssrf. A caller who can set these headers can supply an internal or metadata-service URL and cause the server to send requests to that destination, bypassing the incomplete CVE-2026-27826 remediation. This issue is fixed in version 0.22.0.

Exploitation Scenario

An attacker with network access to a shared mcp-atlassian HTTP deployment, or one who can influence the headers sent by an agent client, sets X-Atlassian-Jira-Url to http://169.254.169.254/latest/meta-data/iam/security-credentials/. The server builds a fetcher using that URL without SSRF validation and sends the request from its own network position. The response, which includes temporary cloud credentials, is returned through the tool output or error messages. The attacker then uses the credentials to pivot into the cloud account, or scans internal services the same way.

Weaknesses (CWE)

CWE-918Server-Side Request Forgery (SSRF)Primary CWE-918Server-Side Request Forgery

CWE-918 — Server-Side Request Forgery (SSRF): The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination.

[Implementation, Operation] If there is a known list of applications, IP, hostname, and/or domains to which the product is allowed to communicate, then use allowlists to ensure that outgoing requests are restricted to these known entities. The allowlists could be in the application code or possibly enforced by firewalls or network segmentation [REF-1511].
[Implementation, Operation] If the product is expected to be able to connect to arbitrary IP addresses or domains, consider using a denylist that includes private network addresses, internal domain names, etc. The denylist could be in the application code or possibly enforced by firewalls or network segmentation [REF-1511].

Source: MITRE CWE corpus.