bkit-rules_skill

Overview

This skill enforces core rules for bkit plugin to ensure consistent AI-native development. It applies PDCA guidance, level detection, agent auto-triggering, and code quality standards automatically. Use it to standardize feature work, fixes, and implementation tasks across project contexts.

How this skill works

The skill inspects project files and CLAUDE.md for explicit declarations, infers project level from folder structure and config files, and classifies incoming requests by size and intent. Based on detection it auto-selects the appropriate agent, enforces PDCA templates and pre-coding checks, and applies code-quality rules such as DRY and SRP. It also blocks or skips automatic actions for documentation-only or research requests.

When to use it

• When developing or implementing features
• When fixing or refactoring code
• When requesting automated agent execution for design, QA, or analysis
• When you need PDCA-driven design or gap analysis
• When enforcing code quality before merging

Best practices

• Start by checking CLAUDE.md and relevant area-specific docs before coding
• Follow PDCA: plan, design, implement, check, act — use provided templates
• Search for existing implementations and reuse utilities before creating new code
• Prefer small, testable functions (SRP) and extract common logic when reused
• Avoid hardcoded values; use meaningful constants and configuration

Example use cases

• Add a new API endpoint: detect level, create design if missing, invoke bkend-expert or enterprise-expert agent
• Fix a bug: compare code and design, run bkit:code-analyzer for targeted review
• Refactor a module: generate plan, update design doc, then execute refactor with checks
• Perform gap analysis after feature implementation: invoke bkit:gap-detector and suggest next PDCA steps
• Run a CI/CD or pipeline review for enterprise projects: auto-trigger bkit:pipeline-guide