llamaindex-development
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides templates for building Retrieval-Augmented Generation (RAG) applications that ingest data from external sources such as local directories and PDF files using components like SimpleDirectoryReader and PDFReader.\n
- Ingestion points: Document loading logic in SKILL.md using SimpleDirectoryReader for local file ingestion.\n
- Boundary markers: The provided code examples do not include explicit delimiters or instructions to the LLM to ignore potentially malicious instructions embedded in the ingested data.\n
- Capability inventory: The skill demonstrates the creation of agents and query engines capable of searching documents and codebases, then generating responses based on that data.\n
- Sanitization: No sanitization, validation, or filtering of ingested content is demonstrated in the provided code snippets.
Audit Metadata