greptimedb-quickstart

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: Recommends downloading and running the GreptimeDB installation script from the official vendor repository.- [EXTERNAL_DOWNLOADS]: Instructs the agent to suggest installing the vendor's MCP server (greptimedb-mcp-server) and adding further skills via npx.- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from database queries and documentation fetches, which presents an attack surface for indirect instructions.
  • Ingestion points: Results from SQL and PromQL queries, table descriptions, and documentation content from docs.greptime.com.
  • Boundary markers: There are no explicit instructions for the agent to treat data from these sources as untrusted or to use specific delimiters.
  • Capability inventory: The skill enables the agent to suggest shell command execution and modifications to the environment's MCP configuration.
  • Sanitization: No data sanitization or validation protocols are defined for the ingested external information.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 12:57 PM
Security Audit — agent-trust-hub — greptimedb-quickstart