habit-hooks-prompting

To stop AI coding agents from silencing linter warnings, you can use targeted coaching prompts structured with the ROSE pattern to nudge the agent toward implementing correct, spirit-aligned fixes for code violations rather than addressing symptoms.

The ROSE pattern for prompt engineering is a structure for creating short, outcome-focused coaching prompts organized around Risk, Observation, Solution, and Expected Outcome to guide AI agents in fixing root causes of code quality issues.

You write coaching prompts for linter, knip, or jscpd violations by generating short, outcome-focused guidance using the ROSE pattern, directing the AI agent to refactor the underlying code quality issue instead of producing shallow or incorrect fixes.

Yes, you can create project-specific override prompts for team coding standards, as the approach supports both bundled prompts for generic linter rules and project-local override prompts to enforce team-specific coding guidance.

The best way to prompt an AI agent to refactor a function with too many parameters is to write a ROSE-pattern coaching prompt that redirects it to split the function into smaller, focused units instead of renaming variables to bypass the rule.

You should use custom coaching prompts when an AI agent's default response to linter, knip, or jscpd rules is to silence warnings rather than fix root causes, leading to persistent technical debt and wasted context window usage.