The Agent Skills Directory

[CREDENTIALS_UNSAFE] (SAFE): Python and SQL code examples use generic placeholders for API keys and database credentials (e.g., 'your-api-key', 'user:pass'). No real secrets or sensitive hardcoded credentials were detected.\n- [EXTERNAL_DOWNLOADS] (SAFE): Installation instructions use official package managers (pip, Docker) for reputable libraries like Qdrant, Weaviate, and Pinecone. These are standard dependencies for RAG development and are sourced from established registries.\n- [COMMAND_EXECUTION] (SAFE): While the skill requests Bash and Write permissions, these are necessary for its intended purpose of implementing and testing machine learning pipelines. No evidence of dangerous shell commands or privilege escalation was found.\n- [PROMPT_INJECTION] (SAFE): No malicious instructions designed to bypass safety filters or extract system prompts were detected. The skill includes architectural advice on using system prompts to constrain LLM responses to provided context, which is a key security measure for RAG systems.

faion-rag-engineer