usenix-security-symposium
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [SAFE]: No security risks, malicious patterns, or obfuscated content were detected within the skill's instructions.
- [EXTERNAL_DOWNLOADS]: The skill directs users to the official USENIX Association website (usenix.org) to verify submission rules. This is a well-known professional organization for computing systems.
- [DATA_EXFILTRATION]: The skill has a surface for indirect prompt injection as it processes user manuscripts. 1. Ingestion points: Manuscript content provided in the author's context (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: None (no tools for network exfiltration, file system modification, or command execution are invoked). 4. Sanitization: Absent. Given the lack of dangerous capabilities, this surface poses no significant risk.
Audit Metadata