seedance
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities detected across all analyzed files. The content is purely instructional and focused on prompt engineering for creative tasks.
- [NO_CODE]: The skill consists entirely of Markdown-based instructions, guidelines, and prompt templates. No executable scripts, binaries, or system-level commands are present.
- [PROMPT_INJECTION]: The skill uses strong instructional language (e.g., "最重要的事", "核心规则") to guide the AI's output and persona, but these do not attempt to override safety filters or extract system-level secrets. They are functional for the intended task.
- [EXTERNAL_DOWNLOADS]: No external network operations (curl, wget, etc.) or remote code fetches were detected. All resources are local to the skill package.
- [CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or sensitive environment variables were found. The skill follows best practices by mentioning .env files for user configuration in the documentation.
Audit Metadata