CVE-2026-45644: XSS in Microsoft Live Share Canvas SDK Poses Insider Threat Risk

Technical summary

CVE-2026-45644 is a stored or reflected XSS vulnerability (CWE-79) in the Microsoft Live Share Canvas SDK stemming from insufficient input sanitization during web page generation. The SDK fails to properly neutralize user-supplied input before rendering it in the DOM, allowing an authenticated attacker to inject malicious HTML and JavaScript. The attack surface is limited to authorized users with valid credentials, reducing the immediate threat landscape but creating insider risk. The vulnerability affects the rendering pipeline where user-controlled data flows directly into HTML output without adequate encoding or filtering.

Business impact

This vulnerability poses a meaningful insider threat to development teams using Microsoft Live Share Canvas. Compromised developers or malicious insiders can inject code into collaborative sessions, potentially leading to source code theft, injection of backdoors into shared projects, lateral movement within engineering environments, or exfiltration of sensitive design documents and credentials shared during pair-programming sessions. Given that Live Share Canvas is used in professional development workflows with access to proprietary codebases, the business impact extends beyond the immediate user to the integrity of software supply chains and intellectual property.

Affected systems

Microsoft Live Share Canvas SDK is affected. The vulnerability applies to all deployments and versions of the SDK until patches are released by Microsoft. Organizations using Live Share Canvas as an extension or integrated feature within Visual Studio Code, Visual Studio, or other IDEs supported by the SDK should assume their installations are at risk if they have not applied vendor patches. The scope is limited to users with valid Microsoft accounts and SDK authentication credentials.

Exploitability

Exploitation requires an authorized user account with access to Live Share Canvas sessions, making this a relatively low-probability but high-impact attack vector. The attacker must also convince or trick a victim into viewing the malicious payload (UI requirement), which slightly reduces exploitability. However, in collaborative team environments where trust is high and users frequently accept and review shared content, this barrier is lower than it would be for unauthenticated XSS. The network-accessible nature of the SDK and the low complexity of the attack (simple input injection) mean that any authenticated user with basic web knowledge could craft an exploit. The vulnerability is not listed on CISA's Known Exploited Vulnerabilities (KEV) catalog, suggesting limited public exploitation as of the publication date, though this should not be interpreted as absence of threat.

Remediation

Microsoft will release patched versions of the Live Share Canvas SDK. Organizations should monitor Microsoft's security advisories and apply updates as soon as they become available through standard SDK distribution channels (NuGet, npm, or vendor download portals). Interim mitigation strategies include restricting Live Share Canvas usage to trusted team members, disabling the SDK in high-security environments until patches are confirmed, implementing Content Security Policy (CSP) headers to limit JavaScript execution scope, and educating users about suspicious collaborative session invitations or unusual content in shared documents.

Patch guidance

Verify the availability of patched SDK versions through Microsoft's official security update channels and your organization's dependency management tools. For Visual Studio Code users, ensure the Live Share extension is updated to the latest version. For development teams using NuGet or npm, check the package registry for security releases and update project manifests accordingly. Organizations should establish a testing window in non-production environments before rolling out updates to development teams, as SDK updates may occasionally affect extension functionality or require configuration changes. Track the patch release date and apply updates within your organization's critical patch SLA—given the HIGH severity rating and insider threat vector, this should be expedited.

Detection guidance

Monitor for suspicious input patterns in Live Share Canvas session logs, particularly HTML tags, JavaScript event handlers, and encoded payloads in user-submitted content. Implement logging at the IDE extension level to capture collaborative session activities and data exchanges. Search for indicators such as <script>, onerror=, onload=, or other event handler injection patterns in session transcripts. Endpoint detection and response (EDR) tools should flag unusual process execution or credential theft attempts originating from development environments. User behavior analytics can identify when authenticated developers suddenly access or exfiltrate unusual code repositories or project files during or immediately after Live Share sessions.

Why prioritize this

This vulnerability merits HIGH priority due to its CVSS 8.0 score, insider threat potential, and the sensitive nature of development environments. While exploitation requires authentication and user interaction, the consequences—code injection, IP theft, supply chain compromise—are severe. Development teams are high-value targets, and the presence of valid credentials significantly lowers the barrier to exploitation. Prioritize patching in organizations where Live Share Canvas is widely used and where development teams handle proprietary or security-critical code.

Risk score, explained

The CVSS 3.1 score of 8.0 (HIGH) reflects a network-accessible vulnerability with low attack complexity and significant confidentiality, integrity, and authenticity impact. The score is tempered by the requirement for valid credentials (PR:L) and user interaction (UI:R), which prevent unauthenticated mass exploitation. However, in the context of insider threats and development environments where credentials are abundant and user trust is high, this score appropriately captures the practical risk. The attack scope is unchanged (the attacker does not break out of the SDK's isolation), preventing a critical rating, but the impact on confidentiality and integrity of collaborative work justifies the HIGH severity.

Frequently asked questions