/digest

Digest Skill

You will create one output from the following web article URL: $ARGUMENTS

• An elegant HTML slide presentation ([article-name]-digest.html)

Follow the four phases below in order.

---

Checkpoint System

Before starting, check for existing checkpoints in .readweave/[article-slug]/ (derive slug from the article title — lowercase, kebab-case, drop articles).

Read the checkpoint reference: Use the Glob tool to find **/checkpoints.md within this plugin's installation directory, then read it for the full format specification.

Resume logic — check in reverse order:

1. If annotations.md exists → load it and skip to Phase 4 (HTML generation)
2. If selection.md exists → load it and skip to Phase 3 (annotations)
3. If extraction.md exists → load it and skip to Phase 2 (selection)
4. If no checkpoints → start from Phase 1

If resuming, report: "Found checkpoint for [phase]. Resuming from [next phase]..."

If the user's input contains --fresh, ignore all checkpoints and start from scratch.

Save after each phase — write the checkpoint file in the format specified by the checkpoint reference before proceeding to the next phase.

---

Phase 1: Fetch & Extract

Fetch the article:

1. Use WebFetch to retrieve the article content from the provided URL
2. If the page is behind a paywall or returns an error, inform the user and stop
3. Extract the article's title, author, publication date (if available), and source/publication name

Parse the content:

• Strip navigation, ads, sidebars, and other non-content elements
• Preserve the article's structure: headings, subheadings, paragraphs, block quotes
• Note any images or diagrams that are central to the argument (describe them in annotations later)
• If the article is very long (5000+ words), mentally divide it into logical sections for systematic extraction

Extract candidate passages:

Collect 8-20 candidate passages depending on article length. For a typical essay (1500-4000 words), aim for 10-15 candidates.

For each candidate passage, capture:

• The exact text, preserving the author's wording precisely
• The section or heading where it appears (if any)
• A one-line note on why this passage matters

What to look for — the five passage types:

• Thesis crystallizations: Where the author states their core claim with maximum clarity. In essays, these often appear 2-3 paragraphs in, not in the opening line.
• Key arguments: Where the author makes a logical move the rest of the piece depends on — a distinction, an analogy, a reductio.
• Striking examples: Where an abstract idea is made concrete through a case, story, or analogy. In blog essays, the example often is the argument.
• Rhetorical high points: Where the prose reaches peak intensity — passages that are beautifully or powerfully said.
• Provocations: Where the author challenges assumptions or received wisdom. The passages that make you stop and think.

What NOT to extract:

• Opening hooks or throat-clearing ("I've been thinking about X lately...")
• Promotional content, calls to action, or self-references to other posts
• Repetition — if the same idea appears twice, take the stronger formulation
• Fragmentary sentences that depend on uncaptured context

Checkpoint: Save extraction.md to .readweave/[article-slug]/ with all candidate passages, article metadata, and section structure.

---

Phase 2: Analyze & Select

Map the argument:

1. Identify the article's core thesis in one sentence
2. Map its structure — how does the argument build? What is the logical progression?
3. Note what the author is responding to — what assumption, trend, or opposing view prompted this piece?

Select excerpts:

From the candidate pool, select 8-15 excerpts for the presentation.

Two-pass selection:

Pass 1 — The essentials. The thesis statement, the key argument, the defining example. These are the 4-6 passages you cannot leave out. If you removed them, the article's argument would collapse.

Pass 2 — The texture. Add passages that give the piece its distinctive voice — the memorable analogy, the sharp provocation, the honest qualification. These make the presentation feel alive rather than merely correct.

Selection quality tests:

1. The idea test: Can you state what this passage captures in one sentence?
2. The stand-alone test: Can someone who hasn't read the article understand this passage?
3. The redundancy test: Does another selected excerpt already capture this idea?

Passage length:

• Sweet spot: 40-150 words / 2-6 sentences. Articles are denser than books — excerpts should be tighter.
• Never truncate mid-thought. Use <span class="elide">[…]</span> for minimal trims.
• Preserve the author's exact wording.

Checkpoint: Save selection.md to .readweave/[article-slug]/ with the argument map, selected excerpts, and critical context.

---

Phase 3: Write Margin Annotations

For each selected excerpt, write 2-3 margin notes.

Labels must be:

• Descriptive concepts: "The Selection Pressure", "Convexity Argument", "Status Game", "Founder Mode"
• Bad: "Key Point!", "Important!", "Note", "Interesting!"

Note body must be:

• Analytical, not summary — never restate what the excerpt says
• Contextual — connect to broader ideas, other thinkers, the author's other work, real-world implications
• Hidden implications — surface what a casual reader might miss
• Concrete — use specific references, named examples, real events

Prose voice:

• Write in a calm, literate register — the tone of a thoughtful reader's pencil notes in the margin, not a lecturer or enthusiast
• Favor observations that unfold naturally over declarations that announce themselves
• Avoid rhetorical questions used for emphasis, exclamation marks, and breathless discovery language ("This is the crucial move", "Here we see the key insight", "This changes everything")
• Let connections speak for themselves — state them plainly rather than dramatizing them
• The best margin notes sound like something a well-read friend might write in pencil — brief, specific, genuinely clarifying, never performing cleverness

This voice also applies to all commentary prose: overview slides and discussion questions.

Quality test: Cover the excerpt text. Can someone learn something new from just the margin note? If not, rewrite it.

Checkpoint: Save annotations.md to .readweave/[article-slug]/ with all annotated excerpts, overview prose, and discussion questions.

---

Phase 4: Generate HTML Slides

1. Read the template: Use the Glob tool to find **/gist-template.html within this plugin's installation directory, then read it with the Read tool. This file contains the full HTML/CSS/JS scaffold for the slide presentation.
2. Read the design guide: Use the Glob tool to find **/read-design-guide.md within this plugin's installation directory, then read it. It documents the HTML structure for each slide type.
3. Replace placeholders:
   • {{BOOK_TITLE}} → the article's title (used in both <title> and sidebar header)
   • {{SIDEBAR_SUBTITLE}} → "Author Name, Source/Date"
4. Populate slides inside the <div class="slides-wrapper">:

Slide structure:

• Slide 0 — Title: class="slide active no-margin" — article title, author, source, date, and an opening quotation from the piece. Include begin-hint.

• Slide 1 — Overview: class="slide no-margin" — the article's core argument in brief. Use a visual diagram (concept-cards, or flow diagram) plus commentary. Frame what follows: "The following pages present [Author]'s argument in [their] own words."

• Slides 2 through N — Excerpts: class="slide" — direct article excerpts with margin annotations. This is the bulk (8-15 slides). Each slide has:

  • page-chapter label with the section heading (if any) or a descriptive label
  • page-divider
  • One or more .excerpt paragraphs (first excerpt on most slides gets .drop-cap)
  • page-source citation (article title, author)
  • .margin div with 2-3 .margin-note elements

• Slide N+1 — Discussion: class="slide no-margin" — 4-6 thought-provoking questions using .discussion-q elements

• Slide N+2 — Closing: class="slide no-margin" — the article's thesis distilled + a final quotation + author/title/source attribution

Formatting rules:

• Use <span class="elide">[…]</span> for elisions within excerpts
• Use literal Unicode: — for em-dashes, ' / ' for single quotes, " / " for double quotes (do NOT use HTML entities — they break in CSS content: properties)
• Apply drop-cap class to the first .excerpt on most excerpt slides
• Each slide needs a descriptive data-title attribute for the sidebar TOC

Slide count guidance:

• Short article (<2000 words): 8-12 total slides
• Medium article (2000-5000 words): 12-18 total slides
• Long article (5000+ words): 18-25 total slides

Save as [article-name]-digest.html in the working directory. Derive [article-name] from the article title: lowercase, kebab-case, drop articles.

---

Important Notes

• Direct excerpts only: The presentation shows the author's actual words. Margin notes provide the analysis.
• Single-file HTML: Completely self-contained. All CSS/JS inline. Only external dependency is Google Fonts CDN.
• Preserve the aesthetic: Reuse the book-page design from the template — parchment colors, Garamond fonts, Caveat margin notes, drop caps.
• Attribute the source: Always include the original URL, author name, and publication date in the title slide and closing slide.
• The margin notes are the value-add: Anyone can read an article. The margin notes — analytical, contextual, concrete — are what make this a digest rather than a bookmark.