The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements a defensive instruction in SKILL.md to treat analyzed content strictly as data and to ignore any text-based instructions embedded within the designs. This proactively mitigates risks from malicious content in the files being audited.
[COMMAND_EXECUTION]: The skill utilizes two local Python scripts (scripts/contraste.py and scripts/escala.py) to perform UI/UX calculations. Analysis of these scripts shows they are limited to the Python standard library and do not engage in network activity, file manipulation, or sensitive data access.
[EXTERNAL_DOWNLOADS]: All external references in skill-card.md and SKILL.md point to well-known, trusted documentation sources (W3C for WCAG guidelines) or the author's own public repository on GitHub, following standard design and development practices.

gestalt