The Agent Skills Directory

[SAFE]: The skill functions as an educational resource for Python AI/ML integration. It demonstrates standard and recommended usage of well-known libraries such as LangChain, LlamaIndex, and Transformers. No malicious patterns, obfuscation, or unauthorized access attempts were detected.
[CREDENTIALS_UNSAFE]: The skill correctly identifies the risk of hardcoding API keys in references/pitfalls.md and provides guidance on using environment variables and .env files for secure secret management.
[INDIRECT_PROMPT_INJECTION]: The skill demonstrates RAG (Retrieval-Augmented Generation) patterns that ingest data from local files. It includes boundary markers in prompts (e.g., "Only answer based on the context") to mitigate prompt injection, and the provided examples have no dangerous capabilities that could be exploited through malicious data.

py-ai