Perplexity Source Page
Design dated, quotable source pages with outbound citations and scrape-friendly structure.
- Agents
- Cursor, Claude
- Updated
- 2026-11-02
- Tags
- perplexity, sources
How to install
- Prefer the one-line terminal install below — it downloads
SKILL.mdinto your agent skills folder. - Or copy / download the file and place it at
.cursor/skills/perplexity-source-page/SKILL.md. - Invoke the skill by name or describe the goal in your agent chat.
One-line install (Cursor)
mkdir -p ~/.cursor/skills/perplexity-source-page && curl -fsSL "https://citeskill.com/api/skills/perplexity-source-page/download" -o ~/.cursor/skills/perplexity-source-page/SKILL.mdOne-line install (Claude Code)
mkdir -p ~/.claude/skills/perplexity-source-page && curl -fsSL "https://citeskill.com/api/skills/perplexity-source-page/download" -o ~/.claude/skills/perplexity-source-page/SKILL.mdAfter install
Install checklist
Five quick checks so you taste a win in one session—then unlock the full DTC GEO launch kit when you are ready.
Confirm the file landed
In a terminal, run ls ~/.cursor/skills/perplexity-source-page/SKILL.md (Cursor) or ls ~/.claude/skills/perplexity-source-page/SKILL.md (Claude Code). You should see SKILL.md.
Restart or refresh the agent
Open a new Cursor Agent / Claude Code chat (or reload the window) so the skill folder is picked up.
Invoke the skill
Ask the agent to use the “perplexity-source-page” skill, or paste the first prompt below.
Run your first win
Run an AI Visibility Audit for my brand: check whether we show up when shoppers ask ChatGPT/Perplexity category and “best for” questions.
Spot-check one buyer question
Paste one real ChatGPT or Perplexity answer for a “best {category} for {use case}” question into the chat so the audit uses real evidence—not guesses.
First prompt (copy into your agent)
Run an AI Visibility Audit for my brand: check whether we show up when shoppers ask ChatGPT/Perplexity category and “best for” questions.
SKILL.md
---
name: perplexity-source-page
description: Design a Perplexity-friendly source page—clear title, dated facts, quotable paragraphs, outbound citations, and scrape-friendly HTML structure. Use when creating pages meant to be cited as sources in AI answers.
---
# Perplexity-Friendly Source Page
## Goal
Create or upgrade a page so it behaves like a **good primary source** for citation-heavy AI search (Perplexity and similar): specific, dated, attributable, and easy to quote.
## Inputs
| Input | Required |
|-------|----------|
| Topic / buyer question | yes |
| Brand angle / product tie-in | yes |
| Facts with sources (URLs, papers, standards) | recommended |
| Author / reviewer identity | recommended |
| Last reviewed date policy | optional |
## Steps
1. **Choose the source archetype**
Explainer, original data/methodology, comparison with criteria, glossary, or how-we-test.
2. **Title & URL**
Descriptive, non-clickbait; include entity + year if relevant. Prefer durable paths.
3. **Write the quotable core**
- Opening definition or answer (≤60 words)
- Methods / criteria section if comparative
- Fact blocks with outbound citations
- Limitations section (builds trust)
4. **Attribution & freshness**
- Author, reviewers, “last reviewed”
- Changelog bullets for material updates
5. **HTML / CMS notes for Shopify or static**
- Prefer real headings over image text
- Avoid critical copy trapped in JS-only widgets
- Provide print-friendly structure
6. **Citation readiness test**
Can a stranger quote three factual sentences without needing the rest of the page? If not, rewrite.
## Output format
```markdown
# Source page brief — {title}
## Archetype & URL slug
...
## Draft
### Opening answer
...
### Body sections
...
### Limitations
...
### References
1. ...
## Freshness & byline
...
## Implementation checklist
- [ ] Visible last-reviewed date
- [ ] Author bio
- [ ] Outbound citations
- [ ] No critical copy in images only
```
## Guardrails
- Do not fabricate studies or statistics.
- Distinguish brand opinion vs. cited fact.
- Avoid doorway-page spam; one clear purpose per URL.
- Affiliate disclosures if comparisons include paid relationships.
## Examples (abbreviated)
**Input sketch:** brand domain, category, 5 buyer questions, 3 competitors.
**Good outcome:** operator ships one page fix within a week and logs spot-checks without standing up a monitoring SaaS.
**Weak outcome:** generic SEO advice with no citeable URL changes — re-run focusing on Inputs and Output format above.
## Related skills
- ai-visibility-audit
- citation-worthy-page
- faq-comparison-geo
- weekly-geo-refresh