skills/brycewang-stanford/awesome-journal-skills/usenix-symposium-on-networked-systems-design-and-implementation/Gen Agent Trust Hub
usenix-symposium-on-networked-systems-design-and-implementation
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override system instructions or bypass safety filters were detected. The instructions are focused on providing domain-specific guidance for academic paper submissions.
- [DATA_EXPOSURE_&_EXFILTRATION]: No sensitive file paths, credentials, or exfiltration patterns were identified. The skill references official academic domains (usenix.org) and internal resource paths consistent with a research environment.
- [OBFUSCATION]: The content was analyzed for Base64 encoding, zero-width characters, homoglyphs, and hidden text patterns (acrostics). No obfuscation techniques were found.
- [REMOTE_CODE_EXECUTION]: There are no commands related to downloading or executing external scripts, nor any installation of unverified packages.
- [INDIRECT_PROMPT_INJECTION]: While the skill is designed to process external data (manuscripts), it lacks the capabilities (such as network writes or system command execution) that would make it vulnerable to indirect injection attacks. It serves as a static analysis tool for conference fit.
- [COMMAND_EXECUTION]: The skill does not use any shell command execution patterns or dynamic context injection syntax.
Audit Metadata