From rrc-R-tools
Behavioral preference for how Claude handles errors and debugging. TRIGGER: whenever the user shares an error message, stack trace, unexpected output, or describes something not working. This includes runtime errors, build failures, type mismatches, test failures, and any 'it broke' scenario. Apply this skill even when the fix seems obvious.
How this skill is triggered — by the user, by Claude, or both
Slash command
/rrc-R-tools:diagnose-before-fixingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
When the user shares an error or something unexpected, follow this sequence:
When the user shares an error or something unexpected, follow this sequence:
The instinct to jump straight to fixing is strong, especially when the cause is clear. But the user may have context you don't — maybe the fix should go somewhere else, maybe the error reveals a deeper issue they want to rethink, or maybe the underlying data or environment is about to change and the fix you'd apply would be wrong. A 30-second explanation costs almost nothing; an unwanted edit can cost real time to understand and revert.
This applies even when:
The only exception is if the user has explicitly said something like "just fix it" or "go ahead and fix errors as you find them" — in that case, you have standing permission to act without pausing.
npx claudepluginhub rrc-associates/claude-rrc-plugins --plugin rrc-R-toolsGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.