llamaindex-development

Pass

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious patterns, obfuscated code, or security bypass instructions were detected within the skill content.
  • [EXTERNAL_DOWNLOADS]: The skill references standard LlamaIndex ecosystem packages (e.g., llama-index, chromadb) intended for installation through official Python package registries.
  • [PROMPT_INJECTION]: The skill describes common RAG architectures that ingest data from local directories. While this introduces an indirect prompt injection surface, it is a standard characteristic of the technology described.
  • Ingestion points: SimpleDirectoryReader is used to load data from the ./data directory (SKILL.md).
  • Boundary markers: Code snippets do not explicitly include boundary delimiters or instructions for the LLM to ignore potentially malicious content within indexed documents.
  • Capability inventory: The documented query engines and agents are designed to process and act upon the contents of retrieved documents.
  • Sanitization: The examples do not demonstrate explicit sanitization of the input data before indexing or querying.
Audit Metadata
Risk Level
SAFE
Analyzed
May 6, 2026, 09:03 AM
Security Audit — agent-trust-hub — llamaindex-development