Walk a user through getting a Google Knowledge Panel for their name — generates customized JSON-LD schema, llms.txt, robots.txt, and a Wikidata creation checklist with their identity filled in. Use when the user wants a Google Knowledge Panel, wants to "looksmax" their search results, asks how to get their face in the box on the right side of Google search, wants structured data for their personal site, or needs help with Wikidata for their own entity.
Scanned 9/6/2026
Install to Claude Code
Skills Directorynpx -y skills add bensig/knowledge-panel-playbook --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of knowledge-panel-playbook?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/bensig-knowledge-panel-playbook)More formats (shields.io, HTML) on the badges page.
---
name: knowledge-panel-playbook
description: Walk a user through getting a Google Knowledge Panel for their name — generates customized JSON-LD schema, llms.txt, robots.txt, and a Wikidata creation checklist with their identity filled in. Use when the user wants a Google Knowledge Panel, wants to "looksmax" their search results, asks how to get their face in the box on the right side of Google search, wants structured data for their personal site, or needs help with Wikidata for their own entity.
---
# Knowledge Panel Playbook Skill
You are about to help a user set up the structured-data foundation for a Google Knowledge Panel.
**Source of authority:** the playbook at <https://benobi.one/posts/google-knowledge-panel-guide/>. Every step in this skill is validated by Ben Sigman's own panel going live in June 2026 after ~10 weeks.
## Phase 0 — Reality check
Before generating anything, confirm the user understands:
1. **This takes ~5–10 weeks.** Schema + Wikidata are done in a day. Google's panel-creation latency is the real wait.
2. **They need a Wikidata-acceptable identity.** Generally means: a real name, a real occupation, and at least one independent press mention or a published book / notable software project. If they have none of those, suggest they wait until they do.
3. **Most "Knowledge Panel services" are scams.** This skill produces the same artifacts those services charge $500–$5,000 for.
If they don't pass step 2's bar, be honest. Don't generate slop.
## Phase 1 — Identity intake
Ask these in batches (don't waste turns on one question at a time). Use AskUserQuestion if available.
**Block 1 — basics:**
- Full legal name (and any common short form, e.g. "Ben" for Benjamin)
- City, state, country of residence
- Citizenship
- Date of birth (optional — they can skip if they prefer it private)
- Personal homepage URL (must exist before this works)
- Headshot — do they have one ready to upload to Wikimedia Commons under CC-BY-SA?
**Block 2 — occupation + entities:**
- Primary occupation (entrepreneur, author, software engineer, etc. — pick from Wikidata occupation Q-items)
- Companies they founded or lead, with URLs
- Notable works: books (with ISBN), software (with repo URL), other creative output
- Co-authors / co-founders by name (so we can check if they have Wikidata items too)
**Block 3 — proof:**
- Press URLs that name them directly (the more the better — minimum 1, ideally 3+)
- Wikipedia article? (rare — usually no)
- Existing Wikidata item? (search wikidata.org for their name to check)
- Profile URLs: LinkedIn, X, GitHub, Crunchbase, Amazon Author, Instagram, YouTube, IMDb, ORCID, Goodreads — whichever apply
**Block 4 — preferences:**
- Education (school + grad year) — only if they have a press source confirming it
- Public DOB / birthplace, yes or no
- Preferred description for Wikidata (e.g. "American entrepreneur, author, and Bitcoin advocate")
## Phase 2 — Generate JSON-LD
Using the template in `templates/jsonld-person.html` (or generate from scratch following the same structure):
- `@graph` containing: Person, all books as `Book`, all software as `SoftwareApplication`, plus `WebSite`
- Link entities via `@id` references
- `Person.sameAs` = array of all profile URLs they gave you, PLUS the Wikidata item URL once they create it
- `Person.affiliation` = `OrganizationRole` with `roleName` + nested `Organization` for each company
- Structured `PostalAddress` for `homeLocation`
- Include `nationality`, `gender`, `alumniOf` (if confirmed), `knowsAbout` (5–8 topics)
Output the complete JSON-LD as a code block ready for them to paste into their site's `<head>`.
## Phase 3 — Generate llms.txt and robots.txt
Drop in templates from `templates/llms.txt` and `templates/robots.txt`, filling in their name + URL.
## Phase 4 — Wikidata creation walkthrough
**Critical rule to repeat at the start of this phase:**
> **Every statement on Wikidata needs a reference URL (P854) or it gets deleted.**
Walk them through creating their Q-item at <https://www.wikidata.org/wiki/Special:NewItem>:
1. Label = full name
2. Description = the short form from Phase 1 Block 4
3. Aliases = nicknames, alternate spellings
Then enumerate statements to add (use the table in `templates/wikidata-checklist.md`), suggesting a reference URL for each from the press URLs they gave you. The table covers: P31, P21, P106, P27, P735, P734, P69 + P582 qualifier, P551, P856, P800, identifier properties (P2002, P2037, P6634, P345, P4862, P2087, P2003, P2397, P2963, P496, P214, P18 image).
For P18 (image) — explicitly walk them through Wikimedia Commons upload at <https://commons.wikimedia.org/wiki/Special:UploadWizard> with CC-BY-SA-4.0 + "It is entirely my own work" before adding the filename to P18.
For each `notable work (P800)` — if it's a book, walk them through creating a separate book Q-item first with `instance of (P31) = book` or `written work`, `author (P50)`, `ISBN-13 (P212)`, `language (P407)`, `publication date (P577)`, then link from their person item.
## Phase 5 — Cross-platform consistency check
Print a checklist:
- [ ] Site schema `homeLocation` city matches Wikidata `P551`
- [ ] Site schema `jobTitle` matches Wikidata `P106` occupation labels
- [ ] LinkedIn headline + location + about all match canonical bio
- [ ] X bio + location match
- [ ] Crunchbase person page matches
- [ ] Amazon Author Central bio matches
- [ ] about.me bio + location field matches
## Phase 6 — Final reminders
1. **After every schema deploy, purge the CDN.** Cloudflare → Caching → Purge Everything. Fastly stale-edge pockets are real.
2. **Validate before you wait:** <https://validator.schema.org/> should show 0 errors / 0 warnings.
3. **Search Console → URL Inspection → Request Indexing** on the homepage after schema changes.
4. **Don't tweak anything for 5+ weeks.** Let Google's crawl cycle catch up.
5. **Don't create org Q-items without press.** They get nominated for deletion fast.
6. **Don't draft a Wikipedia article with an LLM.** Wikipedia AfC reviewers in 2026 actively detect and decline LLM-drafted articles.
## What NOT to do
- ❌ Don't claim this is guaranteed. Google's panel-creation is opaque. Most setups that follow this skill produce a panel within 5–10 weeks; some don't.
- ❌ Don't generate fake press URLs.
- ❌ Don't invent occupations, schools, or biographical facts the user didn't confirm.
- ❌ Don't auto-create the Wikidata item — the user must do that themselves (it's their account, their identity).
## Skill complete
When you reach the end, point the user to:
- The slide presentation: <https://benobi.one/panel.html>
- The full guide: <https://benobi.one/posts/google-knowledge-panel-guide/>
- The repo (where this skill lives): <https://github.com/bensig/knowledge-panel-playbook>
Tell them to come back and report when their panel appears so they can claim it (search their name → bottom of panel → "Claim this knowledge panel" → verify via X or LinkedIn).
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!
