The Agent Skills Directory

[SAFE]: Analysis of the skill instructions and reference materials confirms they are focused on architectural best practices and implementation guidance. No malicious patterns, obfuscation, or unauthorized access attempts were found.- [DATA_EXFILTRATION]: The skill actively provides patterns to prevent data exfiltration of sensitive information. The removePII function in reference/privacy-consent.md explicitly filters out email, phone, name, address, and IP fields before event tracking.- [INDIRECT_PROMPT_INJECTION]: While the skill defines surface areas for data ingestion, it includes mandatory data quality and validation steps. The use of Zod for schema validation in reference/data-quality.md and the requirement for explicit consent markers mitigate risks associated with untrusted data inputs.- [EXTERNAL_DOWNLOADS]: The skill recommends and provides code for established analytics platforms including GA4, Amplitude, Mixpanel, and PostHog. All references are to official SDKs and reputable services.

pulse