The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches the chromadb and requests Python packages from the standard public registry.
[EXTERNAL_DOWNLOADS]: References the official Qdrant Go client library from its GitHub repository for database interactions.
[COMMAND_EXECUTION]: Provides instructions to run database services using Docker and local CLI tools to set up the development environment.
[PROMPT_INJECTION]: The skill contains a Retrieval-Augmented Generation (RAG) implementation pattern in SKILL.md that interpolates external data into model prompts, representing an indirect injection surface.
Ingestion points: Document content retrieved from the vector database via collection.query is used as context.
Boundary markers: Employs standard text labels ("Context:", "Question:") to separate retrieved data from instructions.
Capability inventory: The skill uses bash, read, and write tools, which are common for data management tasks.
Sanitization: The example code does not include explicit sanitization of the retrieved document content before it is processed by the model.

vector-db