The Agent Skills Directory

[DATA_INTEGRITY]: The skill provides detailed guidance on avoiding common data traps in SEO reporting, such as incomplete Search Console data (last 3-4 days) and seasonal demand conflation. These are best practices for accurate data interpretation.
[REMOTE_CODE_EXECUTION]: The skill contains Python snippets for date alignment and YoY comparisons. These are local data processing utilities and do not involve remote script execution or fetching untrusted payloads.
[CREDENTIALS_SAFE]: The skill explicitly advises against hardcoding credentials in the instructions, directing the agent to use environment-specific configuration files (like CLAUDE.md) for deployment credentials.
[COMMAND_EXECUTION]: The skill mentions performing redirect audits which involves crawling site URLs. It provides guidance on handling WAF blocks and rate limiting (429 errors) during the audit process, which is standard for SEO tools.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as search queries and analytics metrics. While this presents a theoretical attack surface, the instructions include logic for normalization and filtering, and the risk is considered low within the context of generating marketing reports.

website-launch-impact-report