The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data to generate prompts and HTML content, creating a surface for potential injection.
Ingestion points: The skill collects a user's name, profession, and background text through the collect phase in SKILL.md.
Boundary markers: The instructions lack specific delimiters or warnings for the agent to ignore instructions embedded in the user-provided data.
Capability inventory: The skill utilizes image_edit tools, file writing (creating index.html from template.html), and shell commands (magick, cwebp) as seen in SKILL.md.
Sanitization: No explicit sanitization logic is provided for user input before it is used in the HTML template or image generation calls.
[COMMAND_EXECUTION]: The skill instructs the agent to execute local command-line tools for image manipulation tasks.
Evidence: SKILL.md contains instructions to use magick for cropping and cwebp for image optimization (e.g., cwebp -q 86 "$SRC" -o "public/ranks/{id}/{file}").

create-bianzu-image