計画、レビュー、レポートをインタラクティブなHTMLページに変換し、ユーザーがその場でアンカー付きのフィードバックやマークアップを行えるようにします。
- GitHubスター
- 8
- ファイル
- 27
- 最終コミット
- 2026年10月28日
- ライセンス
- MIT
- 許可されたツール
- —
render skillは何をしますか
計画、レビュー、レポートをインタラクティブなHTMLページに変換し、ユーザーがその場でアンカー付きのフィードバックやマークアップを行えるようにします。
skillのSKILL.mdから生成された要約。
render skill について
このSKILLは、複雑な計画、レビュー、レポートをインタラクティブなHTMLページに変換することを可能にします。線形なMarkdownの枠を超えて、構造化された視覚的なフィードバックを提供する必要がある開発者やプロジェクトリーダー向けに設計されています。長いドキュメントをスクロールする代わりに、要件マトリックス、依存関係計画、コードレビューボードなどの専用レイアウトを使用して、各項目が特定のコンテンツにアンカーされた状態で操作できます。
ユーザーがドキュメントの視覚化を求めたとき、計画を適切にレビューしたいとき、あるいは他のツールの出力をラップしたいときに、エージェントはこのSKILLをロードすべきです。スラッシュコマンド(例:`render /review-code`)を使用して既存のSKILLをラップしたり、パスを指定して既存のファイルを変換したりできます。このSKILLは、状態を埋め込んだ単一の自己完結型HTMLファイルを生成するため、ユーザーはマークアップやコメントを行うことができ、エージェントは後でそれらを読み取ってアクションを実行できます。すべてのページが機能的かつテーマに沿った状態を維持できるよう、特定のデザインテンプレートとアセンブリスクリプトに依存しています。
skill の SKILL.md から作成されています。すべてのユースケースは、その引用と照合されています。
render skillのインストール方法
各コマンドはskillフォルダをクライアントのskillsディレクトリにコピーします。読み込むには、クライアントを再起動するか新しいセッションを開始してください。プラグインにrender skillをインストールする
/plugin marketplace add mgiovani/cc-arsenal /plugin install cc-arsenal@cc-arsenal-marketplace
Claude Codeにrender skillをインストールする
~/.claude/skills/render/SKILL.md
$ npx degit mgiovani/cc-arsenal/skills/render ~/.claude/skills/renderClaude Code (project)にrender skillをインストールする
.claude/skills/render/SKILL.md
$ npx degit mgiovani/cc-arsenal/skills/render .claude/skills/renderCodexにrender skillをインストールする
~/.agents/skills/render/SKILL.md
$ npx degit mgiovani/cc-arsenal/skills/render ~/.agents/skills/renderGemini CLIにrender skillをインストールする
~/.gemini/skills/render/SKILL.md
$ npx degit mgiovani/cc-arsenal/skills/render ~/.gemini/skills/renderClaude.aiにrender skillをインストールする
npx degit mgiovani/cc-arsenal/skills/render render && zip -r render.zip render # claude.ai → Customize → Skills → + → Create skill → Upload a skill → render.zip
自動チェック
実行日 2026年11月15日完全なレポート →自動チェックに合格しました。これはセキュリティ監査ではありません。エージェントに実行させる前にスクリプトを読んでください。 チェックの仕組み · 点数を上げるための改善点
- ソースリポジトリがリンクされており、アクセス可能mgiovani/cc-arsenal/skills/render is public and reachablegithub.com/mgiovani/cc-arsenal/tree/HEAD/skills/render ↗合格
- オープンソースライセンス (OSI)MITspdx.org/licenses/MIT.html ↗合格
- 過去1年以内にコミットまたはリリースありLast commit 2026-10-28github.com/mgiovani/cc-arsenal/commits ↗合格
- スクリプトにダウンロード実行や認証情報の窃取がない3 scripts read; no download-and-run, remote eval, base64 execution or credential upload合格
- SKILL.md に隠れた指示がないNo override phrases, safeguard switches, credential requests or hidden charactersgithub.com/mgiovani/cc-arsenal/blob/HEAD/skills/render/SKILL ↗合格
- 事前許可ツールが限定されているNo pre-approved tools; the agent asks before using onegithub.com/mgiovani/cc-arsenal/blob/HEAD/skills/render/SKILL ↗合格
- SKILL.md が Agent Skills の形式に従っているthe description does not say when to use the skillgithub.com/mgiovani/cc-arsenal/blob/HEAD/skills/render/SKILL ↗検出事項
SKILL.md
GitHubで表示Render
Build a page the user can work, not a document they have to scroll. Long findings lists, requirement inventories, option matrices and step plans all lose their shape as linear markdown. This skill gives them one, and gives the user a way to answer back that survives the round trip.
Invocation forms
| Form | Behavior |
|---|---|
render <mode> [subject] |
Build that mode's page for the subject |
render /<skill> [args] |
Run the wrapped skill, then render its output in the matching mode |
render <path> |
Read an existing file and convert it to its matching mode |
render |
Show the mode table and ask which one; never pick one silently |
Modes
| Mode | The page is | Per-item verdicts |
|---|---|---|
prd |
Requirements grouped by family, the evidence behind each | keep / change / drop |
plan |
Ordered steps, their dependencies, the files each touches | approve / rework / cut |
review |
Findings by severity and file, with the code excerpt | fix / won't fix / discuss |
audit |
Whole-repo findings, filterable by area and severity | same triage as review |
compare |
Options against weighted criteria, evidence per cell | pick a winner |
brainstorm |
Idea cards, the tension each one resolves, by theme | shortlist / park / drop |
explain |
The one-line answer, the mechanism, then detail on demand | none, comments only |
map |
Module graph, data flow, entry points | none, comments only |
tour |
A guided walkthrough of the code, one stop per file or function | none, comments only |
timeline |
Dated events for an incident or a piece of history | none, comments only |
diff |
Before/after hunks the reader decides on | accept / revise / reject |
report |
A status and metrics summary | none, comments only |
Each mode's contents are specified in references/<mode>.md. Load only the one
you need, and load it after the mode is settled.
Workflow
1. Resolve the mode
From the first argument:
- A mode name: use it.
- A skill name with a leading slash: run that skill first, then map its output
to a mode using the table in
references/wrapping.md. - A file path: read the file, then pick the mode its content matches. State which mode you picked and why, in one line, before building.
- Nothing: print the mode table above and ask. Do not guess.
2. Gather the real content
The page renders what the run actually produced. Never invent an item to fill a grid, never write placeholder copy, never carry an example from a reference file into a real page. If a section would be empty, the page says it is empty and why.
A wrapped skill supplies its own output, and a file supplies its contents. With
a bare mode, do the work the mode implies before rendering: render review on a
diff runs the review first.
3. Load the design guidance
Announce Using artifact-design to calibrate this page's treatment and load the
artifact-design skill before writing any HTML, via the Skill tool where one
exists. If the page will declare a runtime capability, load
artifact-capabilities the same way.
Both are host skills rather than siblings in this repository, so they are absent on some agents. Where neither can be loaded, say so in one line and apply references/page-kit.md directly: it carries the theming rules, the type and layout floor, and the state contract this skill depends on. The page is still built; it just loses the calibration pass.
4. Build the page
No page is hand-authored from a blank file. assets/page.css and
assets/page.js ship the whole design system and runtime; assets/gallery.html
renders every block once and is the visual catalog to check a block's markup
against; assets/templates/<mode>.html is the composition to start from.
Building means:
- Copy
assets/templates/<mode>.htmlfor the resolved mode straight to the real output path,.cc-arsenal/renders/<mode>-<slug>-<YYYY-MM-DD>.html(see the Document skeleton section of references/page-kit.md for why the move is safe even though the template's../page.cssand../page.jslinks are relative). - Replace the whole
/*SAMPLE*/-markedDATAobject with the real content gathered in step 2. Keep the template's structure and itsRender.anchored()calls. - Compose any section the template doesn't already cover only from the block library in references/blocks.md: prefer a block whenever the content has a shape (a sequence, a comparison, a magnitude, a before/after, a hierarchy) over writing another paragraph.
- Visual first, prose last. Every section leads with a block or a diagram.
Prose is at most about three short paragraphs per section, placed in a
.pairbeside its visual. Prefer facts, bars, steps, flow or a diagram over a paragraph. The shell fills the viewport; never add a max-width. - Never hand-write a new CSS color or a raw length. Every value already comes
from a token in
assets/page.css. A diagram comes fromRender.diagram.<type>(spec), never hand-drawn SVG; see references/diagrams.md for the type catalog. - Run
scripts/assemble.pyon that same path, in place (-opointing at the input), to inlinepage.cssandpage.jsinto the single file the gate checks and the reader receives, and let it gate the result against the design contract. A real page never passes--allow-sample: a leftover/*SAMPLE*/marker is a gate failure, not a warning. - Run
npx impeccable detecton the assembled file (needs Node.js:fnm install --ltsorbrew install node) and fix whatever it flags before delivering.
Anchors and the comment affordance are specified in
references/feedback-loop.md, and every mode
carries them through the templates. Reach for
references/diagrams.md only once the content turns
out to have a shape worth drawing as a graph block.
Six rules hold across every mode:
- Every page is annotatable. Anchors and the comment affordance are not a per-mode feature. A page without them cannot return feedback, which is the point of the skill.
- Real content only. See step 2.
- Theme-correct in all three states. Define the light palette as tokens on
bare
:root, redefine those tokens under@media (prefers-color-scheme: dark)guarded by:not([data-theme="light"]), and redefine them again under:root[data-theme="dark"]. Givebodyan explicit token background. A color whose only definition sits inside a media block renders one theme's text on the other theme's background. - The page is greyscale, including verdicts. Every color on the page is
one of the eight neutral tokens in
assets/page.css. There is no hue anywhere, not even for keep/change/drop. A verdict's decision reads from the pressed segment's position in its fixed-order ink-fill control rather than from a color. A dropped or rejected item strikes its title. Corners are square:border-radius: 0everywhere, including form controls and SVGrx/ry. Shadows are limited to a 1px inset outline. No text on the page sets afont-sizebelow 15px, since hierarchy comes from weight and color rather than from shrinking text. See the Color and Shape sections of references/page-kit.md. - Diagrams are data, not drawings. Never hand-write SVG or coordinates;
call
Render.diagram.<type>and let it lay itself out. Over budget means the content gets split into more than one diagram, never squeezed to fit. See references/diagrams.md. - State drives the DOM, never the reverse. The page renders from its embedded state object. Saving serializes that object, never the live DOM.
5. Deliver and report
Report the output path, the published link if there is one, and whatever count matters for this mode. Then say how to answer back: marks and comments are both kept by the page, so the user presses save and tells you. Keep this to two lines.
6. Read the marks back
When the user says they have marked it, read the page back per
references/feedback-loop.md, then act on what it returns. Group your response
by what they decided, not by page order, and name any comment whose anchor no
longer resolves rather than dropping it.
Worked examples
Bare mode (render compare postgres vs sqlite for the vault store): gather
the real criteria from the project's own constraints and build the matrix with
evidence behind every cell. Publish it, report the link, and wait. The user
picks a winner on the page; you read it back and write the decision up.
Wrapping a skill (render /review-security): run review-security to
completion first. Its output is a severity-ranked list with file:line evidence,
which maps to review mode. Build the triage board from the findings it
actually produced, one anchor per finding.
Converting a file (render docs/plans/auth.md): read it first. It turns out
to be an ordered step list with dependencies, so say "rendering as plan mode"
before building. Do not rewrite the plan's content while converting it.
No argument (render): print the mode table and ask which mode. Build
nothing until that is answered.
Notes
- The wrapped skill never needs to know this skill exists.
render /<skill>runs it unchanged and renders what comes out. - Re-running
renderon the same subject updates the same page when the output path matches. This is the intended way to revise, and preserving the reader's marks across it is a required step, not a side effect: read the existing state block and embed it in the new page, per the re-render rules in references/feedback-loop.md. - Five modes carry no verdict controls:
explain,map,tour,timelineandreport. They still carry anchored comments, which is usually the only feedback those pages need.
このスキルのファイル
27render skillに関する質問
Claude Codeにrender skillをインストールするには
npx degit mgiovani/cc-arsenal/skills/render ~/.claude/skills/renderを実行してください。Claude Codeは~/.claude/skills(すべてのプロジェクト)または.claude/skills(単一プロジェクト)からskillsを読み込み、説明がタスクに一致する場合にskillを使用します。
render skillはどのクライアントで使用できますか
Agent Skills形式を読み込むすべてのクライアントで使用できます: Claude Code、Codex、Gemini CLI、claude.ai(ZIPとしてアップロード)。skill内のスクリプトは、クライアントがコード実行を許可した場合にのみ実行されます。
このSKILLで何をレンダリングできますか?
PRD、計画、レビュー、監査、比較、ブレインストーミング、アーキテクチャマップ、コードウォークスルー、インシデントタイムライン、diff、ステータスレポートなどをインタラクティブなページに変換できます。
レンダリングされたページにどのようにフィードバックを提供しますか?
ページにはアンカー付きのコメント機能と、項目ごとの判定(keep/change/dropなど)が備わっており、その場でマークアップを行って変更を保存できます。
他のSKILLをラップして使用できますか?
はい、render /<skill> [args] という構文を使用することで、別のSKILLを実行し、その出力を即座にレンダリングできます。
ダークモードに対応していますか?
はい、デザインシステムはテーマに対応しており、CSSトークンを使用してライトモードとダークモードの両方のカラースキームをサポートしています。