The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and retrieval of external content for use in AI workflows, which creates a potential surface for indirect prompt injection.
Ingestion points: Document values and metadata are ingested into the vector database using the upsert method as shown in SKILL.md.
Boundary markers: Code examples do not demonstrate the use of delimiters or "ignore" instructions to separate retrieved data from agent instructions.
Capability inventory: The skill provides access to database operations including upsert, query, and delete via the pinecone-client library.
Sanitization: The documentation does not specify sanitization or validation logic for data stored in the index.
[SAFE]: The skill uses official, well-known libraries and services.
The pinecone-client dependency and related integration packages (langchain-pinecone, langchain-openai, llama-index-vector-stores-pinecone) are standard industry tools for vector database interaction.
All external URLs point to official Pinecone service domains (pinecone.io).

pinecone