The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a Python script (scripts/generate_schematic.py) that uses subprocess.run to execute a local AI generation script. This is a functional requirement for its diagram generation feature and uses safe argument passing via a list to avoid shell injection.
[EXTERNAL_DOWNLOADS]: The schematic generation tool (scripts/generate_schematic_ai.py) makes network requests to openrouter.ai. This is a well-known AI service aggregator used for generating and reviewing scientific figures based on user descriptions.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from literature searches (e.g., PubMed, general web) and user-provided observations to generate reports.
Ingestion points: External scientific literature fetched via WebFetch and WebSearch tools, and user-provided experimental data.
Boundary markers: The skill instructions do not explicitly define boundary markers for untrusted external content.
Capability inventory: The skill has the ability to execute shell commands (subprocess.run), compile LaTeX (xelatex), and write files to the local system.
Sanitization: No specific sanitization or filtering of external content is implemented before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill uses Python scripts to interact with external AI APIs and manage an iterative refinement loop for images. This behavior is consistent with the skill's primary purpose of providing enhanced scientific visualization.

hypothesis-generation