// This skill should be used when user wants to learn from their collected references, synthesize patterns across sources, or turn research into frameworks from curated reference collections (Are.na channels, bookmarks, saved articles).
Generate minimal HTML pages to review Claude Code output in a browser. Use when terminal output is hard to read, when reviewing lists/tables/drafts, or when user says "show me", "make this reviewable", "quick view", or "open as webpage". Produces unstyled semantic HTML only.
Master LLM-as-a-Judge evaluation techniques including direct scoring, pairwise comparison, rubric generation, and bias mitigation. Use when building evaluation systems, comparing model outputs, or establishing quality standards for AI-generated content.
Build a design system from scratch using the Design Graph methodology. Use when starting a component library, creating design tokens, establishing typography scales, or structuring a theme. Based on Brent Jackson's (jxnblk) constraint-based systems.
Automatically creates user-facing changelogs from git commits by analyzing commit history, categorizing changes, and transforming technical commits into clear, customer-friendly release notes. Turns hours of manual changelog writing into minutes of automated generation.
Design and evaluate context compression strategies for long-running agent sessions. Use when agents exhaust memory, need to summarize conversation history, or when optimizing tokens-per-task rather than tokens-per-request.
This skill should be used to fix visual hierarchy when everything looks equally important, CTAs don't stand out, or users miss key information. Applies Von Restorff Effect, Serial Position Effect, and Law of Prägnanz.
| name | extract-insights |
| description | This skill should be used when user wants to learn from their collected references, synthesize patterns across sources, or turn research into frameworks from curated reference collections (Are.na channels, bookmarks, saved articles). |
Understand the collection contents.
AUDIT CHECKLIST:
- Total items: [count]
- By type: [links, images, text, attachments]
- Link domains: [group by source - medium.com, specific blogs, etc.]
- Blocked sources: [paywalled, bot-detected]
- High-value targets: [articles > images > text snippets]
For Are.na: Run npx tsx scripts/audit-channel.ts <slug> if available, or fetch via API.
Priority order for insight extraction:
For articles/links:
FETCH STRATEGY:
1. Direct fetch (WebFetch)
2. If blocked or inaccessible:
- Flag for manual review
- Note: "[url] - blocked, needs manual access"
3. If success:
- Extract key insights (not full content)
- Note source for citation
For PDFs and attachments:
PDF STRATEGY:
1. Check file size first (if available from metadata)
- < 5MB: Read directly with Read tool
- 5-20MB: Read first 50 pages only
- > 20MB: Skip, flag for manual review
2. If no size available, attempt read:
- If succeeds: extract insights
- If times out or fails: skip and flag
3. For scanned PDFs (no extractable text):
- Skip, flag as "scanned PDF - needs OCR"
Skip criteria (don't waste time):
For each article, extract:
SOURCE: [title] - [url]
TYPE: [framework / case-study / opinion / tutorial / research]
KEY INSIGHT: [1-2 sentence core idea]
ACTIONABLE: [Yes/No - can this become a process?]
FRAMEWORKS: [Any named methodologies?]
Group insights by theme:
CLUSTERING QUESTIONS:
- What problems do multiple sources address?
- What methodologies appear repeatedly?
- What contradictions exist?
- What's unique vs. common knowledge?
Output format:
THEME: [name]
Sources: [list]
Core insight: [synthesis]
Skill potential: [High/Medium/Low]
For high-potential themes, create actionable output:
SYNTHESIS OPTIONS:
1. Skill - if it's a methodology Claude should auto-apply
2. Prompt - if it's a one-time process user would paste
3. Reference doc - if it's knowledge, not process
4. Discard - if it's common knowledge or too vague
Skill criteria:
INSIGHT EXTRACTION REPORT
COLLECTION: [name/url]
AUDITED: [date]
TOTAL ITEMS: [count]
SUCCESSFULLY READ: [count]
BLOCKED: [count]
THEMES IDENTIFIED:
1. [Theme]: [X sources] - [skill potential]
2. [Theme]: [X sources] - [skill potential]
HIGH-VALUE INSIGHTS:
- [Insight] (from: [source])
- [Insight] (from: [source])
RECOMMENDED ACTIONS:
- Create skill: [name] - based on [theme]
- Create prompt: [name] - for [use case]
- Manual review needed: [blocked URLs]
BLOCKED SOURCES (need manual access):
- [url] - [reason]
SKIPPED (too large or unsupported):
- [filename] - [size] - [reason]
| Problem | Solution |
|---|---|
| Most links blocked | Focus on non-paywalled sources first |
| Large PDFs (>20MB) | Skip and flag - likely books, not articles |
| Scanned PDFs | Skip - no extractable text without OCR |
| No clear themes | Collection may be too diverse - narrow scope |
| All common knowledge | Look for unique combinations, not individual insights |
| Too much content | Prioritize recent items, or items with descriptions |
Stop when sufficient insights are extracted:
Extract what's accessible, flag what's not, move on.