Your Codebase, Your Rules: Beyond Standard Linters

A practical two-hour session on turning code quality expectations and team conventions into automated checks so developers and coding agents get feedback on what belongs in your codebase, not just what compiles.

Code can run correctly and still break your team’s rules: too many branches in one function, business logic in the wrong directory, or calls that bypass an approved API. Through small working demos, we will make these violations visible, adjust complexity thresholds, and implement conventions that no standard linter knows about.

The emphasis is on making your expectations executable. Instead of repeating a convention in every prompt or code review, we will express it as a check that identifies violations and provides concrete feedback for correcting them. We will also distinguish likely bugs from team preferences, and explore why passing every check does not automatically mean good design.

Audience: developers, technical leads, and architects interested in keeping codebases consistent as human and AI-generated contributions grow. No prior experience writing custom checks is required.

Outcomes:

• Distinguish correctness checks, complexity measurements, and team-defined conventions
• Configure a complexity threshold and observe its effect on otherwise unchanged code
• Implement custom rules about where particular functions and dependencies belong
• Enforce application-specific naming conventions beyond a tool’s built-in defaults
• Understand when inspecting syntax is sufficient and when deeper analysis is needed
• Give developers and coding agents actionable feedback rather than repeated reminders
• Recognize what automated checks can enforce—and what still requires judgment

Format: two hours with a short introduction, small working demonstrations, deliberate rule violations, discussion of tradeoffs, and questions and answers.