pinecone-n8n
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill generates n8n workflow configurations that interpolate untrusted user data into sensitive fields without explicit sanitization or boundary markers.
- Ingestion points: User-provided inputs such as
[USER_TOPIC]and[USER_FILE_URLS_ARRAY]are directly placed into thesystemMessageof the AI Agent node and theurlfield of the HTTP Request node. - Boundary markers: The generated prompt templates (e.g.,
You are a helpful assistant. Use the Pinecone Assistant Tool to retrieve data about [USER_TOPIC].) lack delimiters or instructions to ignore potential commands embedded within the data. - Capability inventory: The generated workflows possess high-privilege capabilities including fetching arbitrary external files via HTTP, uploading data to Pinecone indices, and interacting with Large Language Models.
- Sanitization: The skill does not include instructions for the agent or the user to validate or sanitize external content before processing it in the ingestion or chat phases.
Audit Metadata