The Agent Skills Directory

[SAFE]: The skill's instructions and reference files were thoroughly audited, and no malicious patterns, deceptive content, or unauthorized behaviors were detected. The functionality is strictly confined to compliance guidance as documented.- [NO_CODE]: This skill consists exclusively of documentation and templates (SKILL.md, references/*.md). It contains no executable scripts (Python, JavaScript, or shell), binary files, or external package dependencies, which eliminates typical attack vectors related to remote code execution.- [DATA_EXFILTRATION]: There are no commands or code patterns that facilitate network communication (such as curl or wget). All analysis is performed within the local execution environment, ensuring no data is transmitted to external servers.- [CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or private tokens are present. Templates provided in the references use non-functional placeholders (e.g., '[ORGANISATION NAME]') for users to customize their own documentation.- [PROMPT_INJECTION]: No prompt injection or safety bypass patterns were found. While the skill analyzes user-provided data in Workflow 1 and 4 of SKILL.md (Data ingestion surface), it lacks any associated tools or capabilities (Capability inventory: None) to perform malicious actions, and therefore the lack of explicit boundary markers does not present a security risk.

gdpr-compliance