Pilot Knowledge Base Rag Setup

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:41 PM UTC
• Version:
  1.0.0

Overview

This is an instruction-only RAG setup skill whose networked data flows are disclosed and aligned with its stated purpose, but users should review trust and data-handling controls before use.

Install only if you intend to run a multi-host RAG pipeline. Before use, review the dependent Pilot skills, verify peer identities before handshakes, restrict network access to trusted hosts, and avoid ingesting sensitive or regulated documents unless permissions, encryption, access control, audit logging, and retention rules are in place.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The README instructs users to ingest documents from S3 and shared directories and to transmit document content and search results between agents, but it does not warn about handling sensitive or regulated data. In a RAG pipeline, this omission can lead operators to ingest confidential material, expose metadata across nodes, or move data into downstream systems without appropriate consent, classification, or retention controls.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The trust model states that when both sides send a handshake, trust is auto-approved with no manual verification, yet the README provides no warning to verify peer identity before establishing trust. In a distributed multi-agent system, this can enable accidental trust of the wrong host, spoofed peers using similar names, or unauthorized data exchange that exposes documents, embeddings, and query traffic.

Missing User Warnings

Medium

• Confidence: 85% confidence
• Finding: The skill sets up a multi-agent pipeline that transmits document content, embeddings, queries, and search results between hosts and exposes a query endpoint to external clients, but it provides no warning about confidentiality, integrity, or trust boundaries. In a RAG system, these data flows can contain sensitive documents and user queries, so omission of security guidance increases the risk of accidental data exposure, interception, or misuse.

VirusTotal

66/66 vendors flagged this skill as clean.