The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and retrieval of external text data, which creates a surface for indirect prompt injection attacks.
Ingestion points: The skill ingests untrusted content via the documents parameter in collection.add() and collection.update() within SKILL.md and references/integration.md.
Boundary markers: There are no explicit instructions or delimiters defined to help the agent distinguish between its system instructions and the retrieved document content.
Capability inventory: The skill includes file writing via chromadb.PersistentClient and network operations via chromadb.HttpClient.
Sanitization: No sanitization, filtering, or validation of the stored documents is performed before they are presented to the agent.
[EXTERNAL_DOWNLOADS]: The skill downloads dependencies and references documentation from well-known and established sources.
Fetches standard library packages including chromadb and sentence-transformers via pip.
Downloads the official chromadb and @chroma-core/default-embed packages via npm.
References the official project repository at github.com/chroma-core/chroma.

chroma