minimax-h3-comic-drama-video
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a local Python script
videogen.pyto facilitate video generation through the AI Hive API. - [EXTERNAL_DOWNLOADS]: The skill requires the
requestslibrary to be installed, which is a standard and well-known Python package for HTTP requests. - [DATA_EXFILTRATION]: The skill transmits user-supplied prompts and local media files to
ai-hive.iclip.cn. While this involves sending data to an external domain, it is the primary and documented function of the skill. - [SAFE]: The script follows security best practices by implementing
chmod 0600on the configuration file (~/.ai-hive/config.json) to protect the user's API key from other local users. - [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for untrusted data through user-provided prompts and media file paths. It possesses the capability to perform network requests and write to the file system. However, as it is a localized tool for a specific service, this surface is considered a standard operational risk. Findings: 1. Ingestion points: CLI arguments in
videogen.py. 2. Boundary markers: Absent. 3. Capability inventory: File read/write and network requests viarequestsinvideogen.py. 4. Sanitization: Absent.
Audit Metadata