Skip to content

Agent Skill Overview

The dbLinter Agent Skill teaches AI coding agents — such as Claude Code, OpenAI Codex, and GitHub Copilot — how to use the dbLinter CLI on your behalf: choosing check, test, or version, scoping a check run to specific files/directories or to new and changed code only, parsing the resulting SARIF/SonarQube reports, and fixing reported issues without disabling rules just to get a clean report.

It is a single SKILL.md file (plus a short reference for parsing report output), following the open Agent Skills standard: YAML frontmatter (name, description, license) followed by Markdown instructions. Any agent that supports this standard can use it once it is installed into that agent’s skills directory — see Compatibility for which agents that includes.

The skill only teaches an agent how to call the CLI correctly and interpret its output — it does not install, configure, or authenticate the CLI itself. See the CLI Overview for that, and one of the subscription pages for the credentials check/test require.

Source and license (Apache-2.0): skills/dblinter in the dbLinter GitHub repository.

  • Installation — install/update via npx skills, which agent(s) to select, and how to verify it’s working.
  • Usage — what to ask your agent for, and which dblinter command it maps to.
  • Compatibility — which agents support Agent Skills today and how each one reads them.