The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill is restricted to generating and managing Markdown-based design documentation within a specific local directory.
[PROMPT_INJECTION]: The skill processes user-supplied screen names and requirements, as well as existing local files (./design/GLOBAL.md), which constitutes an indirect prompt injection surface. The impact is limited because the skill only generates Markdown files and does not have access to sensitive data or executable tools.
Ingestion points: User input for screen descriptions and names; existing design files in the ./design/ directory.
Boundary markers: Absent. The agent relies on internal context to distinguish between instructions and the data being processed.
Capability inventory: Restricted to writing Markdown files to the local ./design/ directory. No network, subprocess, or privilege escalation capabilities.
Sanitization: The skill does not explicitly perform path sanitization or content filtering on inputs used for file creation.

screen-design