The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install code from an external GitHub repository (content-designer/ux-writing-skill).
Evidence: npx skills add content-designer/ux-writing-skill --skill ux-writing-skill -g -y in SKILL.md.
[COMMAND_EXECUTION]: The skill contains bash commands used for environment detection and software installation.
Evidence: grep -q "^das:" ~/.agents/skills/ux-writing-skill/SKILL.md and [ -e ~/.agents/skills/ux-writing-skill ] in SKILL.md.
[PROMPT_INJECTION]: The skill processes user-provided interface copy which represents a data ingestion surface. While no immediate dangerous execution is chained to this input, it remains a surface for indirect prompt injection.
Ingestion points: User-provided microcopy, error messages, and UI text labels.
Boundary markers: None detected in the provided skill text.
Capability inventory: Shell execution (npx, grep) via the ! command syntax.
Sanitization: No explicit sanitization or filtering of input text is documented.

ux-writing-skill