pgvector-semantic-search
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for building Retrieval-Augmented Generation (RAG) systems. This architecture is susceptible to indirect prompt injection where malicious instructions embedded in stored data could be processed by an LLM upon retrieval.\n
- Ingestion points:
SKILL.mddefines storage for external text content in thecontentscolumn of theitemstable.\n - Boundary markers: The provided SQL snippets and instructions do not include delimiters or instructions to treat retrieved content as untrusted.\n
- Capability inventory: The skill involves retrieving stored text based on vector similarity for use in AI-driven search applications.\n
- Sanitization: No sanitization or validation methods are specified for the external data ingested into or retrieved from the database.\n- [EXTERNAL_DOWNLOADS]: The skill references the official GitHub repositories for
pgvectorandpgvectorscale. These are well-known technology resources, and their inclusion is standard for the described functionality.
Audit Metadata