historical-science-video-prod
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed to generate video materials via an automated Python script. It utilizes standard LangChain libraries to communicate with a language model and handles API credentials securely through environment variables. No evidence of malicious code, data exfiltration, or persistence mechanisms was found.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided input for themes and年代 (era), which are included in prompts sent to the language model. This is an attack surface, but the risk is negligible as the output is restricted to static text and markdown files for the user's reference.
- Ingestion points: Command-line arguments (--theme, --era, --core_conclusion) processed in scripts/generate_video_materials.py.
- Boundary markers: Prompts use clear structured headers (e.g., 【䐻题】) to delimit instructions from user input.
- Capability inventory: The skill has file-writing capabilities to a local output directory.
- Sanitization: No explicit sanitization of user strings is performed, which is typical for content generation tools where the user is the primary consumer.
Audit Metadata