The Agent Skills Directory

[SAFE]: The skill acts as a collection of implementation templates and best practices for semantic search applications with no detected malicious patterns.\n- [CREDENTIALS_UNSAFE]: Coding examples correctly prioritize security by retrieving API keys from environment variables rather than hardcoding them in scripts.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides logic to ingest and chunk external documents for search. While this is a surface for indirect injection, the provided templates include preprocessing sanitization logic and the resulting embeddings are used for vector retrieval rather than direct command execution.

embedding-strategies