Atlassian Rovo Vulnerability Exposes Enterprise Data via One-Click Attack

A one-click prompt injection vulnerability, dubbed 'RovoBlast', was discovered in Atlassian's AI assistant Rovo, allowing attackers to exfiltrate sensitive data from connected services like Jira and Confluence. The attack exploits the rovoChatPrompt parameter, enabling attackers to insert malicious instructions that Rovo executes within a user's session. This vulnerability was demonstrated by Varonis researchers at DEF CON 34 and reported to Atlassian, which deployed a fix on July 8, 2026. The attack requires only a single click on a crafted link, making it particularly dangerous. Rovo's extensive access to various platforms means that attackers could potentially access sensitive data without needing to compromise credentials. Organizations are advised to implement robust input validation and limit Rovo's access to mitigate risks. The vulnerability highlights the importance of securing AI assistants in enterprise environments.

Key Points: • RovoBlast allows one-click data exfiltration from Atlassian's Rovo assistant. • The vulnerability affects multiple platforms, including Jira and Confluence. • Atlassian deployed a fix for the vulnerability on July 8, 2026.