TTP Advisory: WriteOut (Writer AI) | No CVE | Severity: Critical (Cross-Tenant Compromise) | Vendor: Writer | Product: Writer Enterprise AI Platform | Discovered by: Sand Security Research | Status: Patched server-side
What Is the Vulnerability
Sand Security Research discovered a critical session isolation vulnerability in Writer’s enterprise AI platform, dubbed ‘WriteOut’. The flaw existed in Writer’s agent preview feature, where sandbox previews could forward session cookies into attacker-controlled agents. An attacker with no prior access could take over any Writer AI organization via a malicious link. From a compromised account, an attacker could access private chats, documents, agent configurations, private models, connectors, and LLM credentials. Depending on the victim’s role, the attacker could also seize administrative control.
Versions Affected
- Writer Enterprise AI Platform — server-side, patched by Writer
Exploited?
No confirmed exploitation in the wild. Writer and Sand Security both state there is no evidence the vulnerability was ever exploited. Writer patched the issue server-side after responsible disclosure.
Fix
Writer patched the vulnerability server-side. No customer action required.
- No action required: Writer applied the fix to all tenants.
Recommendations
- Review AI platform access: This vulnerability highlights the importance of session isolation in multi-tenant AI platforms.
- Audit agent configurations: Review Writer AI agent configurations for sensitive data exposure.
- Broader lesson: AI agent platforms introduce new attack surfaces — session isolation, cross-tenant boundaries, and agent-to-agent trust models should be reviewed.
References
- The Hacker News: Writer WriteOut
- Sand Security Research Report
Part of the Vulnerability Intelligence series on threat-modeling.com. July 8, 2026 Report.
