The Agent Skills Directory

[SAFE]: The skill's primary purpose is auditing for privacy policy consistency. It instructs the agent to extract specific claims from a site's documentation and compare them against actual technical behavior (cookies, storage, data flows).
[SAFE]: The instructions contain explicit safeguards against providing legal advice or making compliance claims. It mandates that any legal specifics must be looked up at official sources (provided in references/official-sources.md) and cited with access dates, rather than relying on the AI's internal training data.
[SAFE]: The skill manages its dependencies through a 'companion' system (references/companion-methods.md). These companions are other skills within the same ecosystem that provide deeper technical analysis (e.g., cookie-and-storage-audit). The logic for using these companions is transparent and includes fallback instructions for when they are not installed.
[SAFE]: Data processing is limited to auditing existing code and UI elements. The skill does not request sensitive local file access (like SSH keys or AWS credentials) or perform unauthorized network exfiltration. It specifically warns against sending real emails during testing in the email-compliance companion section.
[SAFE]: The references/official-sources.md file contains a curated list of trusted governmental and standards-body domains (e.g., europa.eu, ftc.gov, w3.org). These are used for reference only and do not involve remote code execution or untrusted downloads.

policy-consistency