seedance-director
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill package is composed entirely of Markdown documentation and reference guides (
references/). It does not contain any binary executables, shell scripts, or installation files. - [SAFE]: The skill operates as a pure text-processing system, instructing the AI to transform user input into structured video prompts based on cinematic principles like the 180-degree rule and focal length mapping. No patterns of prompt injection, data exfiltration, or persistence mechanisms were found.
- [REMOTE_CODE_EXECUTION]: While the instructions mention using the
python-docxlibrary to generate Word documents, this is framed as a suggestion for the agent's delivery format rather than a malicious execution or an untrusted download. The skill does not provide a script to install or execute arbitrary code.
Audit Metadata