pinecone
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and retrieval of external data, a core requirement for RAG systems. This creates a surface for indirect prompt injection if retrieved content is used in prompts without appropriate sanitization. 1. Ingestion points: Upsert operations defined in SKILL.md. 2. Boundary markers: None present in the boilerplate code. 3. Capability inventory: Read and write access to the vector database via the pinecone-client SDK. 4. Sanitization: No sanitization logic is demonstrated in the examples.
- [SAFE]: The skill utilizes official client libraries and documentation links. There are no instances of malicious command execution, data exfiltration, or obfuscation. All API key references are clearly marked as placeholders.
Audit Metadata