The Agent Skills Directory

[COMMAND_EXECUTION]: The main entry point scripts/generate_infographic.py uses the subprocess module to execute scripts/generate_infographic_ai.py. This is a standard wrapper pattern for CLI tools, though user input flows between the processes via command-line arguments.
[EXTERNAL_DOWNLOADS]: The script scripts/generate_infographic_ai.py makes network requests to openrouter.ai to communicate with Large Language Models (Gemini 3.1 Pro) and image generation models. It also performs research queries via Perplexity Sonar.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it interpolates external, potentially untrusted data directly into prompts for AI models.
Ingestion points: The user-supplied prompt argument in scripts/generate_infographic.py and the content retrieved from external sources during the research phase in scripts/generate_infographic_ai.py (when the --research flag is used).
Boundary markers: The skill does not use specific delimiters or instructions to the AI models to ignore embedded directives in the user content.
Capability inventory: The skill possesses file writing capabilities for saving generated images and logs, and network access to external aggregator APIs.
Sanitization: There is no evidence of escaping, filtering, or validation performed on the user input or research data before it is interpolated into the generation and review prompts.

infographics