The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill makes network requests to openrouter.ai to utilize the Gemini 3.1 Pro and Flash models for research evaluation and image generation tasks. This is a well-known service and the interaction is handled via standard API calls using a user-provided API key.
[COMMAND_EXECUTION]: The script scripts/generate_schematic.py invokes scripts/generate_schematic_ai.py using subprocess.run. The command is constructed as a list, which prevents shell injection, and serves to decouple the CLI logic from the AI processing components.
[INDIRECT_PROMPT_INJECTION]: The skill ingests research documents for evaluation (Ingestion point: SKILL.md workflow). While it lacks specific delimiters (Boundary markers) or sanitization, the capabilities (Capability inventory: network and file-write) are constrained to the skill's specific scientific purpose.
[SAFE]: The skill implements the ScholarEval framework for academic assessment. No evidence of credential theft, persistence mechanisms, or malicious obfuscation was found during the analysis. The file operations are limited to the user-specified output paths for reports and diagrams.

scholar-evaluation