The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local script scripts/generate_image.py to handle image generation tasks. This is standard behavior for extending agent capabilities using project-specific scripts.
[EXTERNAL_DOWNLOADS]: The skill interacts with the official Google Gemini API at generativelanguage.googleapis.com to generate images. This targets a well-known service and is considered safe for the intended functionality.
[PROMPT_INJECTION]: The instructions include guidance for the agent on how to format prompts, including visual style instructions and specific language requirements (e.g., using Simplified Chinese for labels). These are legitimate functional instructions and do not attempt to bypass safety filters.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes user-provided Markdown content to generate prompts for the Gemini image generation model.
Ingestion points: Article content is read as described in Phase 1 of SKILL.md for structure analysis.
Boundary markers: There are no explicit delimiters or "ignore embedded instructions" warnings used when processing the article text.
Capability inventory: The skill uses scripts/generate_image.py which has the capability to write files to the disk and perform network requests to the Gemini API.
Sanitization: No sanitization or validation of the article content is mentioned before it is interpolated into the prompts for the image generation model.

space-gemini-article-illustration