metabase-cli

Pass

Audited by Gen Agent Trust Hub on Oct 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation guides the user to install the @metabase/cli package from the official NPM registry and add the skill stub from the vendor's repository. These resources originate from the official Metabase ecosystem.
  • [COMMAND_EXECUTION]: The skill relies on the mb command-line tool to perform operations such as database inspection, dashboard management, and authentication. It specifically uses mb skills get to fetch instructional content relevant to the installed version of the tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by instructing the agent to execute a command (mb skills get core) and follow the resulting output as further instructions.
  • Ingestion points: The agent context is updated with output from the mb CLI tool (SKILL.md).
  • Boundary markers: No explicit boundary markers or sanitization instructions are provided to separate the tool output from other instructions.
  • Capability inventory: The skill has access to shell command execution (Bash(mb:*)) and file system operations (Read, Write, Edit).
  • Sanitization: There is no explicit sanitization of the tool's output before it is processed as instructions. However, since the source is the vendor's own local CLI, the risk is consistent with standard tool usage.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 10, 2026, 10:19 PM