The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized activities were detected.\n- [DATA_EXFILTRATION]: The skill reads project-specific markdown files such as BRIEF.md, BRAND.md, and IA.md. This access is confined to the local project context and does not involve sensitive system files, environment variables, or hardcoded credentials.\n- [PROMPT_INJECTION]: The skill includes a mechanism for documenting design rule exceptions in a structured table. This is a functional requirement for design consistency and does not constitute a security-relevant prompt injection or an attempt to bypass agent safety guidelines.

visual-direction