The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates Retrieval-Augmented Generation (RAG) by providing patterns for document storage and retrieval. This creates a standard attack surface where untrusted data stored in the database could potentially influence agent behavior if retrieved and processed without sanitization.
Ingestion points: collection.add and collection.upsert methods in SKILL.md and examples/core.md allow for the ingestion of external document content.
Boundary markers: The provided patterns do not include specific delimiters or warnings to ignore instructions embedded within retrieved data.
Capability inventory: The skill utilizes the chromadb client to perform network requests to a Chroma server and manages document storage.
Sanitization: The examples do not demonstrate explicit sanitization or validation of document content before or after retrieval.
[EXTERNAL_DOWNLOADS]: The documentation recommends installing official packages from the Chroma ecosystem, such as chromadb and various @chroma-core/* embedding provider packages, which are standard dependencies for the described functionality.

api-vector-db-chroma