The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the memi CLI tool to extract design metadata and generate design documentation (e.g., memi design-doc <url>).
[EXTERNAL_DOWNLOADS]: The skill fetches HTML and CSS content from user-provided external URLs to analyze their design tokens and components.
[PROMPT_INJECTION]: The skill processes untrusted external data from processed websites, creating an indirect prompt injection surface where malicious instructions could attempt to influence agent behavior during synthesis.
Ingestion points: Public URLs provided as arguments to the design_doc tool or memi command.
Boundary markers: No specific delimiters or safety instructions are defined in the skill documentation to isolate fetched web content from the agent's system prompt.
Capability inventory: The skill is primarily designed to output a DESIGN.md file; no other high-risk capabilities like local file system writes or unauthorized network access were identified.
Sanitization: The skill instructions do not specify any sanitization or filtering of the fetched web content before it is processed by the agent.

design-extract