AgentSkillsCN

principle-check

在给出技术建议之前，先确保其符合既定原则，并以事实为依据，而非单纯基于假设。在提供架构建议、推荐模式，或引用最佳实践之前使用。在建议真正落地之前，及时发现并避免“反射性地照搬专家经验”的倾向。

SKILL.md
--- frontmatter
name: principle-check
description: Pre-response validation ensuring technical recommendations align with established principles and are evidence-based rather than assumption-based. Use before giving architecture advice, suggesting patterns, or citing best practices. Catches reflexive expertise-parroting before it becomes recommendations.

Principle Check Skill

Purpose

Validates technical recommendations against established principles and memory architecture before articulating them. Catches assumption-based or reflexive best-practice parroting, ensuring recommendations are grounded in contextual investigation.

When to Use

  • •Before giving technical architecture recommendations
  • •Before suggesting "best practices" or design patterns
  • •Before recommending defensive code or error handling
  • •Before proposing backwards compatibility measures
  • •When citing documentation as justification
  • •When feeling excited about "comprehensive" solutions
  • •During Learn activity before proposing learnings
  • •When generalizing from specific context to general pattern

Protocol Reference

For complete instructions, see: {{MEMORY_PATH}}/protocols/principle_check.md

The protocol contains:

  • •Phase 1: Recommendation pause (mental checkpoint)
  • •Phase 2: Memory architecture consultation
  • •Phase 3: Principle alignment check (Archaeological Engineering, Proportional Response, etc.)
  • •Phase 4: Red flag detection
  • •Phase 5: Reformulation or proceed
  • •Integration with collaboration patterns
  • •Example applications for common scenarios

Key Principle: Pause before recommending. Verify recommendations are evidence-based (investigated) rather than assumption-based (reflexive expertise).