byted-kickart-game-replicator
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests untrusted data from user-provided video links, character images, game names, and creative prompts. This content is used to populate API payloads for video generation, creating a surface for Indirect Prompt Injection where malicious instructions embedded in source media or text could attempt to influence the agent or the generation service.
- [DATA_EXFILTRATION]: The skill transmits media assets and user-provided metadata to official vendor API endpoints (volcengineapi.com, volces.com). This behavior is consistent with the skill's stated purpose of video replication using vendor cloud services.
- [EXTERNAL_DOWNLOADS]: The skill downloads media files from user-specified remote URLs to the local environment (/tmp) for processing using aiohttp and standard shell utilities like curl and wget.
- [COMMAND_EXECUTION]: The skill orchestrates workflows by instructing the agent to execute shell commands (wget, curl) and internal Python scripts (submit.py, query.py, avatar.py, refer.py) for asset processing and task management.
Audit Metadata