The Agentic Designer v0.3.0

v0.3.0 - 2026-06-11

The 2026-06-11 clip-update round: ten new field-note sections drawn from the week's agentic-design bookmarks, placed across six chapters. All public artifacts rebuilt — self-contained HTML, paged HTML, PDF (320 pages), and EPUB.

New Content

Prompt-template discipline: role / output / constraints / stop (Chapter 03)
A reusable four-block prompt skeleton — role, output, constraints, stop — with five ready-to-run design templates (component spec, heuristic UX audit, task-flow/IA, token mapping, microcopy) and guidance for folding the skeleton into a skill you already run.

Visual prompting: point, draw, or talk on the live UI (Chapter 12)
Cursor's Design Mode as an evolution of the review-and-iterate loop: annotate the rendered UI and the agent maps the mark back to source code, not a throwaway DOM patch. Links the official Design Mode announcement (cursor.com/blog/design-mode).

Generating ideas directly in HTML instead of walls of text (Chapter 04)
HTML as a divergent-ideation surface — laying out the option space as a scannable page of variations rather than a prose list, with a worked prompt that forces breadth and layout.

html-video: a CapCut-style editing layer on Hyperframes (Chapter 09)
An open-source editing layer on the Hyperframes runtime (github.com/heygen-com/hyperframes) adding paginated preview, frame-level text editing, six auto-detected agent CLIs, and MiniMax-generated narration and music.

TypeUI: one markdown file per component (Chapter 03)
A design-skill platform whose transferable lesson is the storage format: one markdown file per concern, served to multiple agents through one MCP endpoint, because agents read prose more fluently than schema blobs.

Multica: agents as assignable teammates on a task board (Chapter 11)
A persistent-infrastructure orchestration model — agents as first-class task-board entries with squad routing, compounding skills, and cron autopilots — contrasted against the in-session orchestrator/subagent pattern.

Generative UI frameworks: when the agent renders the interface at runtime (Chapter 14)
A future-direction note on AG-UI / runtime generative UI, weighing the reviewability cost: runtime-only interfaces trade diffability and versioning for immediacy.

Self-healing /goal commands and the limits of autonomous plans (Chapter 11)
A community /goal upgrade that tries to detect when an autonomous run has gone off-plan and repair the plan, not just the code — flagged as a direction to watch and test.

Xiaohei illustration skill: a single-character explainer-graphic capability (Chapter 03)
A skill that turns text into hand-drawn explainer graphics narrated by one recurring character, making the point that a skill freezes a complete visual fingerprint into a reusable instruction. Includes a figure generated live in the Xiaohei style for this edition.

SkillVault: a marketplace for discovering and distributing agent skills (Chapter 03)
A cross-tool marketplace for skills, agents, hooks, and rules — framed as a discovery layer ahead of the install-and-review gate, not a substitute for reading a skill's source.

Infrastructure

• Bumped book version from 0.2.0 to 0.3.0; applied across chapters 03, 04, 09, 11, 12, 14.
• PDF (320 pages, 200 embedded figures) and EPUB (25/25 package checks) rebuilt from the updated draft source.
• HTML and paged HTML re-published as self-contained single files (figures embedded as data URIs).
• Added inline text link rendering to the clip-apply pipeline so source references resolve as real links; covered by tests.

Downloads

Format	Pages	Size
PDF	320	51 MB
EPUB	19 chapters	51 MB
Self-contained HTML	—	68 MB
Paged HTML preview	—	69 MB

The Agentic Designer v0.2.0

v0.2.0 - 2026-06-04

New Content

Karpathy's CLAUDE.md: Four Rules, 65 Lines, 220k Stars (Chapter 03, section 03.2.5)
Andrej Karpathy's minimal CLAUDE.md hit number one on GitHub trending with 220,000 stars. This section analyzes why a 65-line file with four rules outperforms comprehensive but unfocused harnesses, and maps each rule to design-specific agent behavior: think before generating, simplicity first, surgical changes, and goal-oriented execution.

Beautiful Feishu Whiteboard (Chapter 03, section 03.4-feishu)
A skill that generates editable SVG diagrams inside Feishu/Lark documents with 35 curated color palettes. Serves as a case study in platform-specific constraint handling: the agent must respect Feishu's renderer limitations (no opacity, no gradients, no blur) and work around known export bugs, producing living documentation rather than static screenshots.

SkillSpector: Security Scanning for Agent Skills (Chapter 03, section 03.5)
NVIDIA's security scanner for agent skills checks 64 vulnerability patterns across 16 categories. Covers risk scoring (0-100 scale), installation, and integration with Claude Code (pre-install hooks), OpenCode (validation scripts), and Codex (CI workflows). Includes three real-world use cases: scanning community skills before install, auditing existing team skill libraries, and enforcing team security thresholds in CI.

anydesign: Extracting Design Tokens from Visual Sources (Chapter 04, section 04.4a)
A design skill that extracts structured design tokens from screenshots, live websites via MCP, and Figma files. Produces CSS custom properties, Tailwind config, and JSON token files with confidence levels for each extracted value. Includes companion scripts for drift detection and cross-source verification.

frame.md: Design System Templates for Video (Chapter 09, section 09.5a)
A Hyperframes feature that bridges design systems and video output. Ships 8 templates, skeleton layouts, animation patterns, and shader transitions as markdown files. The agent reads the design spec, the frame.md template translates it, and Hyperframes renders the result.

From Reference Video to Animation (Chapter 09, section 09.6)
A new section covering the video-to-animation pipeline: how agents analyze reference video using ffmpeg frame extraction and LLM motion understanding, AnimSpec's video-to-prompt extraction service (16 output formats), MagicPath's agent-to-canvas handoff for editable designs, and how all three fit into the Remotion/Hyperframes rendering stack. Includes a practical workflow for recreating a competitor's onboarding animation.

Web Access: Firecrawl and Exa MCP (Chapter 10, section 10.8)
Two MCP servers that give agents live web access. Firecrawl provides structured web extraction (search, scrape, crawl, interact, monitor) at scale. Exa provides quick ad-hoc web searches inside agent sessions. Covers when to use which, six practical workflows (competitive audit, design cloning, multi-state UI capture, documentation verification, competitor monitoring, reference collection), and security considerations.

McKinsey 24-Hour Sprint Framework (Chapter 14, section 14.2)
McKinsey's framework for agentic software delivery at organizational scale. Covers the daily cadence (agent-written code by morning, human review by afternoon), eliminating handoffs through shared context, and building knowledge infrastructure (living docs, searchable context, ADRs). Includes a prediction for the 2027-2028 timeline.

Updated Content

Dynamic Workflows (Chapter 11, section 11.6)
Updated the trigger word from "workflow" to "ultracode" with explanation of the May 2026 change that eliminated false triggers. Added new subsection "Workflow Patterns from the Field" covering fan-out, adversarial verification, tournament, and loop-until-done patterns.

Infrastructure

• Updated publication date from May 2026 to June 2026 across all outputs.
• Bumped book version from 0.1.1 to 0.2.0.
• Generated new diagrams: SkillSpector pipeline (ch03-fig-new02), video-to-animation pipeline (ch09-fig-new02).
• Regenerated all output artifacts: HTML, paged HTML, PDF (317 pages), and EPUB.

The Agentic Designer v0.1.1

Changelog - The Agentic Designer

v0.1.1 - May 2026

Release Summary

This release adds the Dynamic Workflows chapter update and a set of bookmark-style agentic design field notes, then republishes all public artifacts: HTML, paged HTML, PDF, EPUB, and cover image.

Content Changes

Dynamic Workflows as orchestration for design work. Expanded Chapter 11 with a full Dynamic Workflows treatment that frames workflow scripts as repeatable design operating artifacts. The update covers when to use workflows instead of live chat, how workflow scripts fan out agents and verify findings, a component-library audit example, reusable design-team workflow prompts, detailed design use cases, model/cost guidance, and the official Claude Code Dynamic Workflows documentation as the implementation reference.

TasteSkill as reusable taste context. Added a bookmark explaining TasteSkill-style repositories as reusable containers for taste, standards, and constraints in agentic front-end work. The note frames skills as guardrails that help agents preserve visual judgment across sessions, with a footer link to the actual TasteSkill repository/resource rather than the social post that surfaced it.

Generated motion for concept slides. Added a bookmark on using video-generation models to animate static concept slides without destroying their visual language. The note keeps the original experiment intact: preserve the slide composition, animate existing illustrations subtly, guide attention from one element to the next, and avoid turning the artifact into generic 3D or live-action output.

Markdown comments as agent review infrastructure. Added a bookmark on markdown-native comments and suggested changes as a practical review layer for agent-assisted work. The note connects tools such as Roughdraft-style review surfaces to the book's larger argument: feedback should live where the next agent and the next version of the author can find it, not only inside a chat transcript.

Claude settings as design environment. Added a bookmark that treats Claude configuration as part of the design interface rather than hidden administration. The note explains why memory, project instructions, extended thinking, search, plugins, hooks, and permission settings affect the quality of agentic design output before the first artifact is generated.

Taste skills from design references. Added a bookmark on turning design references into operational skill files. The note explains how reference analysis can become concrete agent rules around spacing, hierarchy, rhythm, density, typography, composition, and failure modes, so the agent receives design judgment rather than vague adjectives.

Anti-slop skills as design guardrails. Added a bookmark on anti-slop skills as reusable safeguards against generic AI interface patterns. The note frames anti-slop work as a design-stack decision: keep the skill only when it improves context, judgment, and revision speed, and leave it as a reference when it only adds novelty.

Figma export as agent context. Added a bookmark on treating Figma export as structured context for agents instead of a static handoff. The note explains how exported frames, component names, tokens, and layout information can reduce translation loss when moving from design artifact to implementation.

Anchored feedback for agentic design review. Added a bookmark on attaching feedback directly to the artifact under review. The note argues that agentic critique works better when comments are anchored to UI elements, files, or sections, because the next revision can preserve the location and intent of the feedback.

Opening prompts set the design ceiling. Added a bookmark on first prompts as the ceiling of an agentic design session. The note explains how the opening brief sets audience, visual standard, constraints, and failure modes before the agent creates the first artifact, making later refinement less dependent on correcting wrong assumptions.

Artifacts

• HTML: the-agentic-designer.html
• Paged HTML Preview: the-agentic-designer-paged.html
• PDF: the-agentic-designer.pdf
• ePub: the-agentic-designer.epub

Validation

• Export validation passed: 69/69 checks.
• Public package sync used publish-repo --skip-preflight after the same export artifacts passed 69/69 checks; duplicate publish preflight was skipped because the local verifier process hung during the public sync.
• PDF and EPUB were regenerated from the same updated draft source.

Known Issues

• None known.

The Agentic Designer v0.1.0

Changelog - The Agentic Designer

v0.1.0 - May 2026

Release Summary

• Initial public repository package for The Agentic Designer.
• Chapters included: 18.
• License: CC BY-NC-SA 4.0.

Artifacts

• Paged HTML Preview: the-agentic-designer-paged.html
• ePub: the-agentic-designer.epub
• HTML: the-agentic-designer.html
• PDF: the-agentic-designer.pdf

Content Changes

• Add high-level content notes for this release.
• List added chapters, removed sections, major rewrites, image updates, and source-date changes.
• Record any AI-tool version changes that affect instructions or screenshots.

Validation

• Publication preflight and image audit should pass before publishing; internal validation reports remain in the book-writer build output, not in this public book repo.
• Confirm all public download links work.
• Confirm final HTML/PDF/ePub assets open correctly.
• Confirm required copyright, license, AI disclosure, trademark notice, and permissions/credits are present.

Known Issues

• List unresolved source-date, screenshot, formatting, EPUB, PDF, or legal-review limitations.
• Use "None known" only after publication preflight and image audit both pass.