AstrBot CVE-2026-6984

Authentication is required with high-privilege dashboard access (PR:H per CVSS vector). The attacker must have legitimate credentials that grant permission to access the Dashboard API and invoke the create_template function. No special configuration is required beyond a default AstrBot installation with the dashboard enabled. The vulnerability does not require user interaction (UI:N), meaning once the attacker submits a malicious template, exploitation is automatic. Attack scope is unchanged (S:U), meaning the impact is limited to the vulnerable component itself. This limits the attack surface to users or accounts with administrative dashboard privileges, making it lower-risk in environments with strict privilege separation but higher-risk if dashboard access is overly permissive.

CVSS 4.7 with network vector and high privilege requirement (PR:H) suggests this requires an authenticated attacker with elevated dashboard access, limiting the immediate threat surface. However, multiple risk signals conflict with this baseline: (1) publicly available exploit code elevates practical risk beyond the 4.7 score, (2) CWE-1336 (template injection) typically enables information disclosure and code execution, potentially understated by the 4.7 rating, (3) CVSS environmental score modifier (E:P) indicates proof-of-concept existence, confirming public discoverability, (4) no vendor response despite early notification suggests lack of immediate patching urgency. EPSS score not provided, but real-world risk is moderate if administrative dashboard access is accessible to semi-trusted users or if privilege escalation exists upstream. The AC:L (low complexity) factor means exploitation requires no special conditions once authenticated. Defender priority depends on: dashboard exposure scope, whether high-privilege accounts are compartmentalized, and if other entry vectors exist to obtain those credentials.

An authenticated attacker with high-privilege dashboard access navigates to the template creation interface and injects malicious template syntax (e.g., Jinja2 or similar engine directives) into the template parameters. Upon submission, the create_template function processes the malicious input through the template engine without sanitization, allowing the attacker to read sensitive files, enumerate environment variables, or execute arbitrary code depending on template engine capabilities. If the dashboard is accessible to semi-trusted users (e.g., support staff), an attacker could escalate privileges or exfiltrate sensitive application data.