tangjiasanshao-perspective
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No malicious override instructions or safety bypasses were detected. The skill uses 'Trigger words' and an 'Agentic Protocol' to define its persona and workflow, which are standard instructional patterns for personality-based agents.
- [DATA_EXFILTRATION]: No sensitive file paths, credential markers, or suspicious network calls are present. The only external link provided is an attribution URL for the 'Nuwa Skill' framework, which is presented transparently in plain text.
- [REMOTE_CODE_EXECUTION]: The skill does not contain any commands that download or execute external scripts. It is purely text and markdown-based.
- [COMMAND_EXECUTION]: There are no subprocess calls, shell commands, or platform-specific execution patterns like the
!commandsyntax. - [OBFUSCATION]: The reference files contain some formatting artifacts (such as truncated brackets or block characters) likely resulting from manual content extraction, but no intentional obfuscation, encoding (Base64/Hex), or hidden Unicode characters were identified.
- [SAFE]: The skill serves its primary purpose as a creative writing guide and persona model. The instructions are focused on literary structure, 'soul ring' systems, and IP management strategies without introducing technical vulnerabilities.
Audit Metadata