The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/generate_infographic.py uses subprocess.run to launch a secondary Python process (scripts/generate_infographic_ai.py). While it avoids using a shell and passes arguments as a list, it accepts user-controlled content as command-line arguments.
[DATA_EXFILTRATION]: The skill provides a --context-image parameter in scripts/generate_infographic_ai.py that reads local files and sends their base64-encoded content to the OpenRouter API. This functionality could be exploited to exfiltrate sensitive local files if the tool is directed to target them.
[INDIRECT_PROMPT_INJECTION]: The research functionality fetches content from the Perplexity API and includes it in generation prompts.
Ingestion points: External research data is ingested in scripts/generate_infographic_ai.py via the Perplexity Sonar Pro model.
Boundary markers: The prompt uses standard headers to delineate research data but does not include explicit instructions to ignore potentially malicious instructions embedded in that data.
Capability inventory: The skill has the ability to execute sibling scripts via subprocess.run and perform network requests to external APIs.
Sanitization: Research data is interpolated into the final generation prompt without sanitization or escaping.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to openrouter.ai and perplexity.ai to access LLM and image generation services. These are well-known technology services and are used for the skill's primary purpose.

infographics