feat(skills): add skill-creator agent instructions (grader, analyzer, comparator)

[JeremyDev87]

Summary

• Add 3 agent instruction files for skill-creator's Eval/Improve/Benchmark modes
  • agents/grader.md — Grades eval outputs against assertions, produces grading.json with expectations[].text/passed/evidence
  • agents/analyzer.md — Analyzes benchmark.json + grading results, outputs strengths/weaknesses/suggestions/priority report
  • agents/comparator.md — Blind comparison of two skill version outputs, outputs {preferred, confidence, reasoning} JSON
• All input/output formats aligned with references/schemas.md JSON schemas
• Each agent is independently executable with no cross-agent references
• Includes Iron Law, Red Flags tables, step-by-step processes, and grading criteria

Test plan

• All 3 files created in packages/rules/.ai-rules/skills/skill-creator/agents/
• Schema alignment verified: grader→grading.json, analyzer→benchmark.json, comparator→custom JSON
• codingbuddy workspace CI: lint, format, typecheck, test:coverage, circular, build — all pass
• rules-validation: ajv-cli agent schema validation + markdownlint — all pass
• EVAL review: Critical=0, High=0

Closes #740

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
codingbuddy-landing	Ready	Preview, Comment	Mar 21, 2026 1:55pm