tarkovsky-perspective
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill uses standard persona instructions to adopt the role of Andrei Tarkovsky. It defines triggers and response rules ('Directly respond as Tarkovsky', 'Use the first person') which are standard for character-based AI behaviors and do not attempt to bypass safety filters or ignore system constraints.
- [DATA_EXFILTRATION]: There are no commands or instructions that access sensitive files, environment variables, or attempt to send data to external servers. The skill operates entirely within the conversational context.
- [REMOTE_CODE_EXECUTION]: No scripts, package installations, or remote downloads are present. The file consists purely of Markdown instructions and structural guidance.
- [COMMAND_EXECUTION]: The skill does not contain any shell commands, subprocess calls, or dynamic execution patterns.
- [EXTERNAL_DOWNLOADS]: The skill contains a single link to a GitHub repository (github.com/alchaincyf/nuwa-skill) as a project attribution. This is an informational link to a well-known service and does not involve automated downloads or execution of untrusted code.
- [OBFUSCATION]: The content is clear and uses standard Markdown. No hidden characters, Base64-encoded payloads, or homoglyph substitutions were detected.
- [PRIVILEGE_ESCALATION]: The skill does not attempt to gain administrative access or modify system configurations.
Audit Metadata