alterlab-scientific-schematics
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/generate_schematic.pyacts as a wrapper, calling the implementation script viasubprocess.run. It uses an argument list rather than a shell string, which is a safe practice that prevents command injection. - [EXTERNAL_DOWNLOADS]: The skill communicates with the OpenRouter API (
openrouter.ai) to utilize image generation and vision models. This targets a well-known service and is documented neutrally. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because user-supplied diagram descriptions are interpolated into model prompts without robust sanitization or boundary markers. • Ingestion points: Diagram descriptions provided as command-line arguments in
scripts/generate_schematic.pyand processed inscripts/generate_schematic_ai.py. • Boundary markers: Absent; the user input is placed directly after a text label in the prompt template. • Capability inventory: The skill has the ability to write files to the local disk, perform network requests to an external API, and execute local Python scripts. • Sanitization: No sanitization is performed on the user input before it is used in the prompt template.
Audit Metadata