The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which represents a vulnerability surface for indirect prompt injection attacks.
Ingestion points: The skill retrieves data from external sources via WebFetch, mcp__magic__21st_magic_component_inspiration, mcp__magic__logo_search, and various Figma MCPs as documented in references/tooling-integration.md.
Boundary markers: There are no explicit instructions or delimiters defined within the skill to isolate external content or warn the agent against executing embedded instructions found in these sources.
Capability inventory: The skill is granted Write and Edit permissions and utilizes the mcp__magic__21st_magic_component_builder tool to generate and output production-ready React and Tailwind code (noted in SKILL.md and references/tooling-integration.md).
Sanitization: The skill lacks defined sanitization, validation, or filtering mechanisms for the data ingested from external APIs or design files.

web-design-expert