Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add laguna821/Achmage-Skills --skill component-consulting-v3git clone --depth 1 https://github.com/laguna821/Achmage-SkillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/laguna821/achmage-skills/component-consulting-v3)<a href="https://agentmods.dev/skills/laguna821/achmage-skills/component-consulting-v3"><img src="https://agentmods.dev/badge/skills/laguna821/achmage-skills/component-consulting-v3/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/laguna821/achmage-skills/component-consulting-v3"><img src="https://agentmods.dev/badge/skills/laguna821/achmage-skills/component-consulting-v3.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00246 | $0.07654 |
| Opus 5 | $0.00123 | $0.03827 |
| Sonnet 5 | $0.00049 | $0.01531 |
| Haiku 4.5 | $0.00025 | $0.00765 |
Grade A, and why
component-consulting-v3 scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 467 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Component Consulting v3
Design.md posture: consumes. Reads a project's
DESIGN.md(or any token file) and produces a component prescription grounded in a corpus of real code. It does not generate the design system — pair with a token-emitter (e.g.design-md-scaffold[[design-md-scaffold]]).
You are a page author who buys components, not a component generator and not a decorator. The governing conviction (안창현, 2026-08-28):
프론트엔드도 글쓰기다. 한 개의 HTML 로딩된 페이지 = 한 편의 완결된 글/책 챕터. 내용이 좋아야 디자인이 비로소 산다. 컴포넌트는 문단과 장면 단위로 독자의 몰입을 위해 배치하는 장치다.
Two mechanical guarantees distinguish v3 from its ancestors:
- The corpus is real.
corpus/holds 4,196 vendored components with actual HTML/CSS/TSX (MIT, commit-pinned, author-attributed) + 79 index-only references. Candidates come from deterministic retrieval (references/retrieval.md), never from live-scan sourcing or model memory. Every STUB in the output is copied fromcode_path— a model-invented stub is a defect, not a fallback. - The page is a text first. Before any component is named, the
page-as-text protocol (
references/page-as-text.md) must produce a Reader profile, a Genre/Experience verdict, and a Beat sequence. No beats → no prescription.
§0 — Governing principles
The one-sentence test (unchanged from v2, still the floor):
"Because the page's thesis is X and this beat must do Y for reader R, this component earns its place."
Hard rules:
- Text before components. Run page-as-text §R1→R3 first. A page with an EMPTY grounding verdict is refused — 소재 없으면 설계하지 않는다 (the design-side twin of ach-writing-system-v7's P0 rule).
- Corpus-first, deterministically. Source every candidate through
references/retrieval.md. Live web access is delta-checking only (new components since the corpus pin), never the sourcing path. - One signature moment, many quiet structures. Exactly one loud,
awwwards-grade moment per page — AND a density floor: ≥ 6
recognizable gallery signatures · ≥ 10 distinct form factors · ≥ 3
interaction layers per page (the
component-density-contractnumbers, now a blocking gate — "one signature" was never "few components"). - License is a gate, not a note. Only
MIT / Apache-2.0 / BSD / ISCcomponents may be prescribed.usable: falserows (unknown / commercial license,ats_verdict: cut) are never recommended — the query tool hides them by default. (v3.5 각주 — F31: 주석은 게이트가 아니다, 게이트는 enum 이다. 큐레이터의 한국어 주석에만 "ATS-banned" 가 있고 enum 이 공란인 행 11개가 기본 질의에 노출됐던 실측이 동인 — annotation-enum 일관성은verify_corpus.py규칙 9가, 슬롭 패턴은scan_slop.py의slop_flags[]가 기계 필드로 강제한다. Step 7 의 슬롭 밴 목록도 이제 소싱 시점에 문다: BAN 급 플래그(gradient-text·auto-advance)는 기본 숨김이다.) - Reskin ≠ erase. The skin becomes the project's tokens; the
component's
preserve_signaturestays observable in the DOM, and the element carriesdata-gallery="{source}:{name}"+data-component. Unrecognizable = zero consumption (laundering).
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- .claude-plugin/plugin.json 1.1 KB
- corpus/index/aliases.json 4.4 KB
- corpus/index/components.jsonl 3734 KB
- corpus/index/corpus-manifest.json 1.6 KB
- corpus/index/legacy-unmatched.jsonl 6.3 KB
- corpus/index/views/by-altitude/atom.jsonl 1120 KB
- corpus/index/views/by-altitude/molecule.jsonl 60 KB
- corpus/index/views/by-altitude/section.jsonl 55 KB
- corpus/index/views/by-intent/i01.jsonl 99 KB
- corpus/index/views/by-intent/i02.jsonl 24 KB
- corpus/index/views/by-intent/i03.jsonl 23 KB
- corpus/index/views/by-intent/i04.jsonl 17 KB
- corpus/index/views/by-intent/i05.jsonl 19 KB
- corpus/index/views/by-intent/i06.jsonl 70 KB
- corpus/index/views/by-intent/i07.jsonl 4.2 KB
- corpus/index/views/by-intent/i08.jsonl 29 KB
- corpus/index/views/by-intent/i09.jsonl 70 KB
- corpus/index/views/by-intent/i10.jsonl 53 KB
- corpus/index/views/by-intent/i11.jsonl 55 KB
- corpus/index/views/by-intent/i12.jsonl 43 KB
- corpus/index/views/by-type/accordion.jsonl 465 B
- corpus/index/views/by-type/alert.jsonl 202 B
- corpus/index/views/by-type/announcement-banner.jsonl 219 B
- corpus/index/views/by-type/auth-form.jsonl 3.0 KB
- corpus/index/views/by-type/avatar.jsonl 1.2 KB
- corpus/index/views/by-type/background-effect.jsonl 226 B
- corpus/index/views/by-type/background-pattern.jsonl 240 B
- corpus/index/views/by-type/badge.jsonl 738 B
- corpus/index/views/by-type/balance-slider.jsonl 212 B
- corpus/index/views/by-type/before-after-slider.jsonl 231 B
- corpus/index/views/by-type/bento-grid.jsonl 228 B
- corpus/index/views/by-type/bento.jsonl 232 B
- corpus/index/views/by-type/breadcrumb.jsonl 216 B
- corpus/index/views/by-type/breadcrumbs.jsonl 240 B
- corpus/index/views/by-type/button.jsonl 365 KB
- corpus/index/views/by-type/calendar-widget.jsonl 216 B
- corpus/index/views/by-type/card-accent.jsonl 224 B
- corpus/index/views/by-type/card-stack.jsonl 205 B
- corpus/index/views/by-type/card.jsonl 215 KB
- corpus/index/views/by-type/carousel.jsonl 924 B
- corpus/index/views/by-type/checkbox.jsonl 51 KB
- corpus/index/views/by-type/code-block.jsonl 517 B
- corpus/index/views/by-type/combobox.jsonl 529 B
- corpus/index/views/by-type/command-palette.jsonl 215 B
- corpus/index/views/by-type/comparator.jsonl 1.0 KB
- corpus/index/views/by-type/confetti.jsonl 214 B
- corpus/index/views/by-type/contact.jsonl 747 B
- corpus/index/views/by-type/content.jsonl 2.7 KB
- corpus/index/views/by-type/counter.jsonl 940 B
- corpus/index/views/by-type/cta.jsonl 3.0 KB
- corpus/index/views/by-type/data-table.jsonl 215 B
- corpus/index/views/by-type/device-frame.jsonl 907 B
- corpus/index/views/by-type/device-mockup.jsonl 228 B
- corpus/index/views/by-type/divider.jsonl 194 B
- corpus/index/views/by-type/dock-nav.jsonl 204 B
- corpus/index/views/by-type/drawer.jsonl 221 B
- corpus/index/views/by-type/dropdown-menu.jsonl 974 B
- corpus/index/views/by-type/effect.jsonl 19 KB
- corpus/index/views/by-type/expandable-card.jsonl 220 B
- corpus/index/views/by-type/faq.jsonl 2.9 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 467 lines · 246 tokens per session scan A 75d73f569079
component-consulting-v3 is a skill published in the GitHub repository laguna821/Achmage-Skills (5 stars, last pushed 11d ago), licensed Apache-2.0. It adds 246 tokens to every session and 7,654 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
webgl-holographic-foil
A self-contained WebGL2 hero: thin-film interference over a crushed-foil surface whose palette shifts with the viewing angle; move the cursor to tilt the film.
html-ppt-hermes-cyber-terminal
OpenDesign + BYOK: choosing and wiring your own model, hands-on — cost, quality, and the routing decision. Built as a decision-grade AI literacy deck for engineers, IT, applied-AI teams.
html-ppt-taste-brutalist
16:9 HTML deck in tactical-telemetry / CRT-terminal taste. Deactivated-CRT charcoal slides, white-phosphor monospace, hazard-red accent, scanline overlay, ASCII syntax, density over decoration. Distilled from Leonxlnx/taste-skill brutalist-skill (Tactical Telemetry mode).
visual-ralph
Visual Ralph orchestration for frontend UI from generated references, static references, or live URL targets, using $ralph with built-in visual verdict and pixel-diff evidence until the implementation matches and leaves a reproducible design system.
accessibility
Consolidated accessibility skill entrypoint for WCAG 2.2, ARIA Authoring Practices, cognitive accessibility, Section 508, EN 301 549, design intent verification, and the Accessibility Planner workflow.
make-resume
A Chinese-language tool for creating editable HTML resumes that can be changed in a browser and printed to PDF. It uses available resume templates when they are installed and otherwise provides a simpler fallback.