The Agent Skills Directory

[PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it retrieves data from external collections and uses an agent to synthesize answers.
Ingestion points: The ask.py and query_search.py scripts utilize the QueryAgent to ingest data from Weaviate collections based on natural language queries.
Boundary markers: The scripts do not implement explicit delimiters or 'ignore embedded instructions' warnings when passing retrieved content to the underlying model.
Capability inventory: The skill includes capabilities to read files, search the filesystem, and perform network operations to the database and external data hubs.
Sanitization: There is no content-level sanitization or filtering of the text retrieved from Weaviate before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The example_data.py script fetches datasets from the Hugging Face Hub, which is a well-known and established service. These datasets (e.g., jamescalam/ai-arxiv2) are used to populate the database with example data for testing and demonstration purposes.

weaviate