skill-architect
Skill Architect
Architect portable, high-quality agent skills from a vendor-neutral perspective. Use this skill as an independent planning, creation, review, refactoring, comparison, distillation, reconciliation, and teaching layer for any agent skill repository or runtime.
Why this skill exists
Several high-signal upstream skills overlap while emphasizing different strengths:
| Source | Strength to preserve | Boundary to avoid |
|---|---|---|
| Anthropic Skill Development | Strong structure, trigger descriptions, progressive disclosure, validation checklist | Claude Code plugin-specific assumptions are not portable as-is |
| Anthropic skill-creator | Skill evals, baseline-vs-with-skill testing, trigger optimization, blind comparison, benchmark loops | Too heavy and Claude-specific to be the whole portable rubric |
| OpenAI skill-creator | Codex/OpenAI compatibility, concise scaffold guidance, agents/openai.yaml metadata |
Too OpenAI-specific to be the central standard |
Treat this skill as a portable synthesis: it directs, plans, and reviews skill work at the architecture level, then routes into source-specific, project-specific, or repository-specific details only when needed. Concrete examples of this rubric in practice can be inspected across the skills4sh skills.
Operating mode
Start with the current working context, not an upstream rubric. Check the CWD, project files, installed skill location, or repository convention first. External rubrics inform content quality, but local rules win on packaging, frontmatter, manifests, lockfiles, CI, and release gates when those rules exist.