The Agent Skills Directory

[SAFE]: The skill defines instructional logic for design workflows and does not contain any malicious code, obfuscation, or unauthorized data access patterns.
[EXTERNAL_DOWNLOADS]: The instructions mention optional partner skills from the uxKero ecosystem and specific community design tools (e.g., emilkowalski/skills). These are presented as informative installation pointers (npx skills add) to be used by the operator or agent and do not involve silent or forced remote code execution.
[INDIRECT_PROMPT_INJECTION]: The skill identifies surfaces for processing project source code, Figma files, and screenshots. While this creates a data ingestion surface, it is necessary for the skill's purpose as a design orchestrator.
Ingestion points: Reads project code (tokens, components, routes), screenshots, and Figma files.
Boundary markers: Absent in the instruction text.
Capability inventory: Writing project-specific skills, delegating to sub-agents, and performing UI audits.
Sanitization: No specific sanitization methods are described for imported visual references.

kero-method