modular-rag-mcp-server

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to clone a project repository from an external, non-whitelisted GitHub account: https://github.com/jerry-ai-dev/MODULAR-RAG-MCP-SERVER.git.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data (PDFs, Markdown) into a RAG pipeline, which can lead to indirect prompt injection attacks.
  • Ingestion points: The IngestionPipeline (referenced in SKILL.md) processes PDF documents and other file types to create embeddings.
  • Boundary markers: There are no explicit instructions or delimiters mentioned to isolate retrieved document content from the system prompt, which may lead to the agent following instructions embedded in the data.
  • Capability inventory: The skill configures an MCP server that executes shell commands and performs network API requests to multiple providers (OpenAI, Anthropic, Cohere, etc.).
  • Sanitization: The provided documentation does not specify any sanitization, filtering, or validation logic for the content of ingested documents.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 10:43 PM
Security Audit — agent-trust-hub — modular-rag-mcp-server