The Agent Skills Directory

[SAFE]: The skill performs legitimate article analysis and image generation tasks using built-in tools (image_gen) and platform-specific commands (cowrite_get_page, cowrite_upload_asset).
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes external Markdown content to generate image prompts.
Ingestion points: External article content is ingested via the cowrite_get_page tool in Phase 1.
Boundary markers: No explicit delimiters or 'ignore embedded instructions' warnings are used when processing the ingested text.
Capability inventory: The skill has access to image_gen for content creation and cowrite_insert_after for modifying the article page.
Sanitization: The skill contains explicit safety instructions to 'never rewrite the source article' and to 'insert only successful final images', which mitigates the impact of potential malicious instructions embedded in articles.

article-batch-illustration