Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Open Orbit briefing skill — selected by the Orbit pipeline when the user has two or more connectors connected. Pulls the past 24 hours of activity from every authenticated connector (GitHub, Linear, Notion, Slack, 飞书, Calendar, Gmail, Drive, Sentry, Vercel, …) and renders a single adaptive bento-grid dashboard at the top of "我的设计". Each connector module picks its own UI form (list, avatar stack, status ring, heatmap, file grid, alert card, …) based on the data shape it returns, so the layout sca
.claude/skills/nexu-io-orbit-general/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 67% | 0% |
Cross-connector morning briefing that lives at the top of "我的设计". Pulls the past 24 hours of activity from every authenticated connector and lays them out as one editorial bento dashboard.
Step 1. Open and read the shipped example.html in this folder before writing any output. That file is the canonical design — your job is to reproduce it, not reinterpret it.
Step 2. Mirror the example's structure 1:1:
<script> block at the end (link injection)Step 3. You may freshen mock data values (counts, names, times) so they read as "today" — but you must not invent new UI elements, sections, modules, badges, callouts, ribbons, banners, decorations or chrome that aren't already in example.html. If a detail is not in the example, it does not belong in your output.
The body sections below are a reference for the visual language and tokens — they are not a license to add features the example doesn't already render.
This skill ships with its own complete visual language baked into example.html. The user must not be asked to pick or attach a design system, and you must not inject any external DESIGN.md tokens into the output.
example.html.
This is a hard constraint: an Orbit briefing must read as Open Orbit's own editorial bento language, not as the user's brand.
--bg: #FAF7F2 /* off-white page */
--surface: #FFFFFF /* card */
--fg: #1A1816 /* ink */
--muted: #6B6660 /* secondary text */
--border: #EAE5DD /* 1px hairline only */
--orange: #D86A47 /* accent (CTAs, hero highlight, meeting blocks) */
--green: #2E7D5B /* ok / done */
--yellow: #C9982E /* waiting */
--red: #C0473A /* alert / fail */
--radius-l: 24px /* outer container */
--radius-m: 16px /* bento cards */
--radius-s: 12px /* inner blocks */Type stack:
'Cormorant', Georgia, serif — KPI numerals, Hero h1,Top 3 serial numbers, italic comment quotes
'Inter', -apple-system, system-ui, sans-seriffont-variant-numeric: tabular-numsNo shadows. No gradients. No emoji as primary visuals. Connector icons must be monochrome line SVG (1.5 stroke).
Left: ☀ 早安, Eli (Cormorant 38px, with , in --orange). Right of name: · 2026 年 5 月 6 日 · 星期三 (muted, 18px). Far right: round avatar (40px) + small ⚙ + ✕ icons.
Each cell: serif number (Cormorant 64px, --fg) over a muted uppercase tracking label (Inter 11px, letter-spacing 0.06em). Optional ▲/▼ delta tag in --green/--red next to the number. Suggested labels: 待办 / 待 review / 会议 / @ 我 / agent 跑完.
Horizontal time axis from 09:00 → 19:00, hour ticks below. Meeting blocks: filled --orange rounded rectangles spanning their start/end, with the meeting name + attendee count inside. Deep-work suggestions: pale-green translucent bands behind the axis. "Now" indicator: a 1px vertical --red line with a pulsing dot (@keyframes pulse 2s ease-in-out infinite) and a tiny 现在 label.
Each card: huge serif numeral 1 / 2 / 3 (Cormorant 96px, in --fg) left-aligned; one-sentence task headline (Inter 18px medium); a meta row at the bottom with the connector source label + line-icon + 等待 Xh waiting time. Cards have --border 1px outline only.
Render 10–16 modules. Sizes vary: data-rich connectors take a 2-column or 2-row span, simple ones stay 1×1. No two modules should look identical. Pick UI per the data family below.
Title left: 5 人在等你 · 最久 22h (serif 24px). Right: 5 overlapping circular avatars (44px, ~8px overlap), each with the person's name + waiting reason underneath in 12px muted.
Left: Open Orbit · auto-generated 06:42 · N connectors. Right: 由 Nexu Labs 出品. Border-top 1px, all text 12px muted.
| Family | Examples | UI form | |---------------|---------------------------------------|------------------------------------------------------| | Code collab | GitHub, GitLab, Bitbucket | Status-dot list (open/merged/closed/CI fail) + reviewer count, optional 2–3 line diff preview | | Task mgmt | Linear, Jira, Asana, ClickUp | Issue list with colored status dot + priority bars; for cycle, add a small ring or progress strip | | Comms | Gmail, Slack, 飞书 IM, Outlook | Round avatar + one-line quote, accent color for "awaiting reply" | | Knowledge | Notion, Confluence, 飞书 Doc | Doc title + 2-line excerpt block; comment quote in italic serif | | Time | Calendar | Already lives in the global timeline; module form: agenda list with start time gutter | | Alerts | Sentry, Datadog, PagerDuty | Big red Cormorant number (e.g. 4), 7 small squares as 7-day heatmap, plus 1 latest error line | | Status | Vercel, GH Actions, Netlify | Colored status dot per recent build/deploy + branch + duration | | Files | Drive, Dropbox, Box | Filename list with tiny thumbnail squares + "edited by" attribution | | Board | Trello, Miro, FigJam | 3 compact kanban columns with rounded card chips | | Finance | Stripe, PayPal, banking, Brex | Cormorant currency number + 7-day sparkline + last 3 transactions list | | CRM / Sales | Salesforce, HubSpot, Pipedrive | 3-column deal pipeline (Open / Negotiation / Won) + 1–2 priority contact cards | | Support | Zendesk, Intercom, Help Scout | Ticket queue list with SLA timer pill (green / yellow / red) + assignee avatar | | Analytics | Google Analytics, Mixpanel, Amplitude | Mini funnel chart (4 bars descending) + 1-line cohort delta (▲ 12% W/W) | | Infrastructure| AWS, GCP, Kubernetes, Docker | Resource meters (CPU / mem / disk percent bars) + last 2 deployment lines | | Security | 1Password, Auth0, Okta | Event list with red shield for high-severity items + audit timestamp | | Voice/Misc | unknown connector | See Fallback heuristics below |
When a connector doesn't match any family above, infer by the data shape it returns:
status field → treat as Task mgmt (status-dot list)from / subject → treat as Comms (avatar + quote)If still ambiguous, fall back to a status-dot list (the safest default).
| Don't | Do | |---|---| | Render every module as the same card shape | Vary by family — Alert = big red number + heatmap; Status = status-dot list; Files = thumbnail grid; Comms = avatar + quote | | Render Sentry / PagerDuty as a plain list | Big red Cormorant number + 7-day heatmap + latest error line (TypeError: …) | | Render Calendar as a plain text agenda | Visualize on the horizontal timeline at the top; module form is an agenda list with start-time gutter | | Use placeholder names like "Service A / Project X" | Infer plausible real names from the connector type — GitHub → nexu-io/open-design, Sentry → frontend-prod, Linear → ENG / DES cycle 24, Stripe → Pro plan / Acme Co. | | Use lorem ipsum filler | Write specific mock copy that reads as a real workday — names, numbers, errors, paths, percentages | | Mix emoji and SVG icons in the same module set | Use monochrome line SVGs (1.5 stroke) consistently for all connector icons; emoji are reserved for hero greeting and section anchors only | | Square or rounded-square avatars | Always circles; sizes 28 / 32 / 40 / 44 px depending on context | | Drop shadows / gradients / glows on cards | Flat surfaces only; differentiate cards with the 1px #EAE5DD hairline border | | Use brand colors from the user's design system | Use exclusively the canvas tokens above (#FAF7F2, #1A1816, #D86A47 …) — Orbit's own editorial language |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 51,991 | 35,475 | -32% | 1 | 1 | 0% | 8,278 | 10,872 | +31% | 0 | 0 | — |
case-02 | fail→fail | 40,390 | 50,731 | +26% | 1 | 1 | 0% | 8,276 | 10,870 | +31% | 0 | 0 | — |
case-03 | fail→fail | 50,090 | 37,089 | -26% | 1 | 1 | 0% | 8,261 | 10,855 | +31% | 0 | 0 | — |
case-04 | pass→pass | 51,590 | 42,632 | -17% | 1 | 1 | 0% | 8,231 | 10,825 | +32% | 0 | 0 | — |
case-05 | pass→fail | 17,750 | 35,067 | +98% | 1 | 1 | 0% | 3,552 | 10,813 | +204% | 0 | 0 | — |
case-06 | pass→fail | 30,665 | 49,910 | +63% | 1 | 1 | 0% | 4,816 | 10,808 | +124% | 0 | 0 | — |
case-07 | fail→pass | 15,965 | 4,623 | -71% | 1 | 1 | 0% | 2,873 | 3,431 | +19% | 0 | 0 | — |
case-08 | fail→pass | 12,557 | 2,960 | -76% | 1 | 1 | 0% | 2,172 | 3,063 | +41% | 0 | 0 | — |
case-09 | fail→pass | 20,000 | 7,901 | -60% | 1 | 1 | 0% | 3,024 | 3,769 | +25% | 0 | 0 | — |
case-10 | fail→pass | 20,377 | 13,959 | -31% | 1 | 1 | 0% | 3,439 | 4,572 | +33% | 0 | 0 | — |
case-11 | fail→pass | 15,592 | 9,233 | -41% | 1 | 1 | 0% | 2,308 | 3,861 | +67% | 0 | 0 | — |
case-12 | fail→pass | 20,252 | 7,255 | -64% | 1 | 1 | 0% | 2,937 | 3,560 | +21% | 0 | 0 | — |
case-13 | fail→pass | 20,805 | 16,215 | -22% | 1 | 1 | 0% | 2,831 | 5,904 | +109% | 0 | 0 | — |
case-14 | fail→fail | 13,684 | 6,220 | -55% | 1 | 1 | 0% | 2,254 | 3,374 | +50% | 0 | 0 | — |
case-15 | fail→pass | 17,273 | 21,844 | +26% | 1 | 1 | 0% | 2,468 | 5,983 | +142% | 0 | 0 | — |
case-16 | fail→fail | 16,301 | 3,105 | -81% | 1 | 1 | 0% | 2,053 | 3,020 | +47% | 0 | 0 | — |
case-17 | fail→pass | 15,917 | 11,858 | -26% | 1 | 1 | 0% | 2,107 | 4,720 | +124% | 0 | 0 | — |
case-18 | fail→pass | 17,152 | 18,451 | +8% | 1 | 1 | 0% | 2,911 | 6,240 | +114% | 0 | 0 | — |
case-19 | fail→fail | 16,799 | 17,472 | +4% | 1 | 1 | 0% | 2,208 | 4,889 | +121% | 0 | 0 | — |
case-20 | fail→pass | 13,328 | 4,378 | -67% | 1 | 1 | 0% | 2,194 | 3,388 | +54% | 0 | 0 | — |
case-21 | fail→pass | 16,386 | 7,299 | -55% | 1 | 1 | 0% | 2,575 | 3,973 | +54% | 0 | 0 | — |
case-22 | fail→fail | 10,669 | 3,155 | -70% | 1 | 1 | 0% | 1,079 | 3,058 | +183% | 0 | 0 | — |
case-23 | pass→pass | 10,054 | 3,673 | -63% | 1 | 1 | 0% | 1,252 | 3,098 | +147% | 0 | 0 | — |
case-24 | fail→pass | 16,146 | 5,109 | -68% | 1 | 1 | 0% | 2,396 | 3,445 | +44% | 0 | 0 | — |
case-25 | fail→fail | 18,080 | 2,281 | -87% | 1 | 1 | 0% | 1,627 | 2,961 | +82% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted. The headline lift of +44 percentage points is the difference between those two pass rates over the 25 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.