SQLite MCP Server
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install packages using
npxandpipfrom the Model Context Protocol organization, which is a well-known service in the agent tooling ecosystem. These references are documented neutrally as they originate from a recognized service. - [INDIRECT_PROMPT_INJECTION]: The skill enables an agent to ingest data from SQLite databases, creating a surface for indirect prompt injection where data could contain instructions that influence the agent.
- Ingestion points: SQLite database records retrieved by the agent from local files.
- Boundary markers: None; the skill does not provide instructions or delimiters to separate database content from agent commands.
- Capability inventory: The agent can execute SQL queries and ingest the resulting data into its context.
- Sanitization: No sanitization of the database content is specified before processing by the agent.
Audit Metadata