The Agent Skills Directory

[DYNAMIC_EXECUTION]: The scripts/check-setup.py script uses the __import__ function to programmatically verify that required Python packages like langchain_core and langchain_openai are installed. While this involves dynamic module loading, the module names are derived from a hardcoded list within the script and do not incorporate untrusted input.
[INDIRECT_PROMPT_INJECTION]: The skill provides architecture for RAG pipelines and agents that process external data, such as web pages fetched via WebBaseLoader in templates/rag-pipeline.py. This creates a surface for indirect prompt injection where untrusted external content could attempt to manipulate the agent's logic. * Ingestion points: WebBaseLoader is used to fetch external documents in templates/rag-pipeline.py. * Boundary markers: The templates demonstrate basic integration without explicit prompt delimiters or 'ignore' instructions for retrieved content. * Capability inventory: The skill enables tool-calling capabilities as shown in templates/agent-with-tools.py. * Sanitization: No explicit sanitization or filtering of retrieved document content is shown in the provided templates.

langchain