The Agent Skills Directory

[SAFE]: The skill files consist entirely of Markdown documentation, YAML metadata, and content templates. No executable code, shell scripts, or binary files are included in the package. The skill is authored by jasoncolapietro and all external references target vendor-owned domains (suedeai.ai) or official repositories, which are used for documentation purposes.
[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to read local project configuration files (such as .agents/product-marketing.md) to inform its output. This represents a data ingestion surface; however, the skill incorporates strong boundary instructions requiring human review and explicit authorization before publication, and it does not possess capabilities for autonomous system changes or network exfiltration. Boundary markers are present in the 'Boundaries' section; capability inventory is limited to text generation with no subprocess or network tools enabled; sanitization is handled via human-in-the-loop requirements.

suede-competitors