← ClaudeAtlas

html-reportinglisted

Render a self-contained HTML report or dashboard (task summaries, code/operational audits) in one canonical dark-terminal shell. NOT for markdown-only output or the mermaid viewer (use mermaid-viewer).
atretyak1985/swarmery · ★ 4 · Data & Documents · score 76
Install: claude install-skill atretyak1985/swarmery
# Purpose Wrap already-authored content in one reusable, **self-contained** HTML shell — inline CSS, no external assets, no build step — so every report an agent emits shares one visual language: a dark terminal theme with severity-coded cards, metric tables, and collapsible sections. Used by the `session-closeout` skill's Phase 8 dashboard and `@code-reviewer`'s Phase 5 audit. Presentation only. This skill never gathers data, measures anything, or computes a number — the caller supplies the content and this skill renders it. # Rules (never violate) 1. Never fabricate a metric or a health score; format only what the caller passed. 2. Keep the file self-contained — no CDN link, no external `src`/`href`, no `<script>` unless explicitly required. 3. Use the canonical shell verbatim and its severity classes; never re-style a report or hand-roll inline colors. 4. Markdown stays the canonical artifact; the HTML is its optional mirror. 5. Respect the body budget — `<main>` ≤ 300 lines (summary) / ≤ 500 (audit); consolidate rather than pad, and never leave a `{{…}}` placeholder. 6. No secrets or PII in the output; verify the written file is non-empty. # Resources - Read `resources/shell.md` when rendering: the canonical HTML shell to copy verbatim plus the content-to-component map. - Read `resources/render-procedure.md` first: the input table, the two section skeletons (summary, audit), the 5-step procedure, the self-check, common mistakes, and a worked example.