The Agent Skills Directory

[SAFE]: The skill implements strong safety guardrails by explicitly prohibiting the handling of real PII and the provision of legal advice. It focuses on procedural workflows and educational reference material.
[COMMAND_EXECUTION]: The script scripts/validate_skill.py is used for local structural validation of the skill's frontmatter and file structure. It uses standard Python libraries and performs no network operations, dynamic code execution, or access to sensitive system files.
[EXTERNAL_DOWNLOADS]: The skill provides links to official regulatory documentation from well-known and trusted sources, including gdpr-info.eu, edpb.europa.eu, ico.org.uk, cppa.ca.gov, and hhs.gov.
[PROMPT_INJECTION]: The instructions include clear 'When to Use' and 'Do NOT Activate For' sections, which act as boundary markers to prevent the agent from performing out-of-scope tasks or interpreting statutes authoritatively.
[DATA_EXPOSURE]: Safety rules mandated in SKILL.md explicitly require the redaction of sensitive data and the use of synthetic values in examples, mitigating risks of accidental data exposure.

privacy-compliance