build-sites-as-specs

Architect a website-generation system that structurally cannot produce slop, by compiling a validated specification into output instead of generating HTML directly. Use when building an AI site generator, a templating/CMS pipeline, a multi-tenant site builder, or any system where a model or template produces web pages and you keep getting hallucinated facts, inconsistent structure, weak SEO/accessibility, and unmaintainable markup. Triggers include "build a website generator", "AI keeps generating slop sites", "how do I make site generation reliable", "spec-driven site building", "stop the model inventing facts on the page".

설치로 이동

저장소: ariaxhan/the-agent-library
최근 소스 활동: 2026년 6월 28일 04:02
감지된 SKILL.md 언어: 영어
스타: 1
포크: 0

설치 방법

기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.

Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.

직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.

npx skills add https://github.com/ariaxhan/the-agent-library --skill build-sites-as-specs

명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.

로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.

소스 파일 검토

설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.

파일 탐색기

2 개 파일

SKILL.md

소스 지침 · 읽기 전용 미리보기

검색으로 돌아가기 설치로 돌아가기

extract-invariants

Turn a borrowed metaphor, analogy, or cross-domain inspiration into a testable mechanism, or prove it is decoration and discard it. Use when a design leans on language from another field ("immune system", "evolution", "market", "quorum", "crystallization"), when someone proposes copying how another domain solves a problem, or when reviewing a system whose names sound deeper than its code. Produces either a named domain-free invariant with evidence and falsifiers, or an explicit rejection with the reason it failed.

2026-07-20

design-loud-failures

Design systems so a failure can never present as success. Use when building or reviewing pipelines, extractors, reconciliations, checkers, or any computation whose output humans act on; when missing inputs, unreadable files, or crashes could pass silently; when a tool "found" an explanation that feels too convenient; or on "make this fail loudly", "could this lie to us", "harden the failure paths". Produces named invariants, explicit skip and no-answer states, honest cause classes, and adversarial regression fixtures.

2026-07-10

debate-decisions

A compact decision protocol for non-trivial choices, reviews, governance calls, and live disagreement. Use implicitly at decision points, even when nobody says "debate", whenever a choice is high-impact, hard to reverse, policy-bearing, or has multiple plausible paths that evidence alone cannot settle. Skip for trivial, cheap-to-reverse, or deterministically decidable choices.

2026-07-07

이 저장소의 모든 Skills 보기

Build Sites as Specs

Most AI site builders go straight from a prompt to HTML. That single shortcut is the source of every recurring failure: hallucinated facts, inconsistent structure, broken accessibility, weak SEO, vendor lock-in, output you can't diff or edit. The fix is to stop treating the website as the artifact. The specification is the artifact. HTML is just one compiled output of it.

Business Data → Brief → SiteSpec → Renderer → HTML

Reframed as a compiler: the Brief is source code, the SiteSpec is the intermediate representation, the design system is the standard library, the Renderer is the backend target, HTML is the compiled output. Once you see it this way, the model stops being the center of the system: it fills slots in a spec, it never writes markup, controls design, or invents facts.

The rule above all others

Do not build the AI part first. Build the grammar (the spec), the compiler (the renderer), and the validator (the policies) so well that the AI becomes the least interesting, most replaceable part of the system.

Build order, inside-out, do not skip ahead:

Specification → Validation → Rendering → Generation

Generation comes last. If a hand-authored spec can't be validated and rendered into a production-quality site offline, with zero network and zero model calls, the architecture isn't done, and adding generation on top will only multiply the cracks.