skills/mizchi/skills/node-sqlite-vec/Gen Agent Trust Hub

node-sqlite-vec

Pass

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill implements a storage and retrieval system for vector search (RAG), which involves ingesting and storing arbitrary text data (body) in a SQLite database. This data could contain malicious instructions designed to influence an AI agent when the data is retrieved and processed.
  • Ingestion points: The add function described in SKILL.md accepts a body string from external sources to be stored in the items table.
  • Boundary markers: There are no explicit instructions or delimiters provided to wrap the retrieved content to prevent an agent from following instructions embedded within the text.
  • Capability inventory: The skill enables file-system writes through the DatabaseSync API for database persistence.
  • Sanitization: While the code correctly uses prepared statements to prevent traditional SQL injection, it does not provide mechanisms to sanitize or filter the stored text content for downstream AI safety.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 12, 2026, 03:23 AM