The Agent Skills Directory

[SAFE]: The skill implements a robust privacy-preserving architecture for handling user profiles. The scripts/profile_store.py utility includes explicit checks to exclude sensitive fields (token, secret, password, etc.) from being persisted, and uses atomic writes to prevent data corruption. All identified network or package operations target official or well-known development resources, such as PyPI for PyYAML, which is standard for the skill's validation logic. The instructions strictly mandate that legal drafts be treated as unverified boilerplate for human review, explicitly forbidding the AI from claiming legal certification or fabricating business facts. Execution is scoped to user-defined directories with no persistent auto-execution or privilege escalation patterns.

lov-legal-pages