The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes technical assets such as HTML and CSS code fetched from external URLs, which serves as a potential vector for indirect prompt injection if the source content contains malicious instructions.
Ingestion points: The skill uses web_fetch to download HTML and screenshot assets from dynamic URLs provided by the Stitch MCP server as defined in SKILL.md.
Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following instructions potentially embedded within the ingested technical code.
Capability inventory: The skill has access to the Write tool to generate files and web_fetch to retrieve external data.
Sanitization: The skill does not perform explicit sanitization or filtering of the fetched HTML before parsing it for design tokens.

design-md