deep-work

Generate dark-themed self-contained HTML learning dashboards and multi-page portals from JSON specs.

Updated Apr 14, 2026
One-click install
npx skills add https://github.com/amitkarpe/agent-skills --skill deep-work-amitkarpe
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deep-work
Source: https://github.com/amitkarpe/agent-skills/tree/main/skills/deep-work
Command: npx skills add https://github.com/amitkarpe/agent-skills --skill deep-work-amitkarpe

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires diagrams, graphviz, pandoc, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? Dense technical topics, architecture reviews, and long research are hard to retain from plain notes; this Skill turns them into durable, offline HTML dashboards and portals designed for long-term learning and study. ## Core Features & Use Cases - Single-page and portal modes: Produce one self-contained dashboard or a multi-page portal with shared dark nav, version watermarks, and an export-safe all-in-one.html convertible to docx via Pandoc. - Deterministic JSON-spec rendering: Write a report.spec.json validated against a component catalog (Hero, Timeline, Glossary, Quiz, Flashcards, CommandBlocks, ArchitectureMap) and render it with Python scripts instead of hand-writing HTML. - Architecture Lens: Add AWS/account-boundary diagrams, swimlanes, and AMI lineage maps using inline SVG or the Python diagrams library with Graphviz. - Use Case: After an AMI factory architecture review, generate a deep-work portal with architecture, inventory, runbook, and export pages, publish it to the local web surface, and convert the all-in-one page to a Word document for offline sharing. ## Quick Start Ask the agent to create a deep-work dashboard that explains a dense technical topic with a concept map, timeline, glossary, and quiz, then publish it as a local HTML page.

Frequently Asked Questions about deep-work

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a deep work learning dashboard from a technical topic?▼

Provide the topic or evidence and request a deep-work dashboard. The Skill builds a self-contained dark HTML page with a hero summary, concept map, timeline, glossary, flashcards, and quiz, then publishes it to the local web surface after URL validation.

How do I render an HTML report from a JSON spec?▼

Write a report.spec.json using only components from the component catalog, run validate_spec.py to check it, then run render_html_from_spec.py to produce the HTML and validate_html.py to verify the output before publishing.

When should I use portal mode instead of a single-page dashboard?▼

Use portal mode when the output spans more than one page or the request mentions a portal or KB. A portal has a shared nav on every page, an index.html entry point, stable filenames, and an all-in-one.html export page.

Can I convert the dashboard to a Word document?▼

Yes, the all-in-one.html page is export-safe: it avoids details/summary tags and JavaScript-toggled content so Pandoc can convert it. Run pandoc all-in-one.html --from html --to docx to produce the Word file.

What diagram options exist for AWS architecture pages?▼

The preferred option is the Python diagrams library with Graphviz to generate PNGs embedded as base64. If those are unavailable, the Skill falls back to inline SVG patterns such as account boundary boxes, swimlanes, and AMI lineage forks.

Why does the dashboard reject external fonts or CDN links?▼

The design requires fully offline, self-contained pages, so publish_html.py blocks external URLs, credential-like strings, and white screen backgrounds outside print CSS. Any page with remote assets fails validation before publishing.