The Agent Skills Directory

[COMMAND_EXECUTION]: The entry point script scripts/generate_infographic.py utilizes subprocess.run to execute a secondary AI-handling script. This is a standard architectural pattern for modular Python applications and does not represent a security risk.
[EXTERNAL_DOWNLOADS]: The skill makes legitimate outbound connections to openrouter.ai to facilitate communication with generative AI models and research tools. These external references are documented and necessary for the skill's functionality.
[CREDENTIALS_UNSAFE]: The implementation handles an OpenRouter API key via environment variables or .env files. This aligns with industry best practices for secure secret management in development tools.
[DATA_EXFILTRATION]: Data provided by the user and information retrieved through the research feature are sent to the OpenRouter API. This data flow is expected and essential for the infographic generation process.
[PROMPT_INJECTION]: The skill contains a potential surface for indirect prompt injection within scripts/generate_infographic_ai.py when the --research flag is enabled.
Ingestion points: Data retrieved from external web sources via the Perplexity research API is interpolated into the generation prompt.
Boundary markers: The research content is placed under clear headers in the prompt but does not use strict structural isolation.
Capability inventory: The skill's primary capabilities are reading/writing files (images and JSON logs) and making network requests to the AI provider.
Sanitization: External data is incorporated into the prompt without specific sanitization filters.

infographics