Name Drop: do AI assistants mention your brand? A GEO check
Name Drop is a skill for Claude that measures whether AI assistants name your brand when people ask about your category, and explains why when they do not. It uses a fixed panel of questions that never contain your name, records who is named instead and which sources are cited, and orders the fixes by evidence.
One of the 14 skills in the pack
Ask it in your own words
- "Is my brand mentioned by ChatGPT?"
- "Why doesn't AI recommend my product?"
- "Compare my AI visibility with these three competitors."
- "my app has the same name as some old software and perplexity keeps mixing us up"
What happens next
- It checks the name and crawler access. It searches your bare name, then reads your robots.txt for the assistants' search agents, which are not the training crawlers.
- It builds the panel. Questions in each market's language: discovery, comparison, problem, decision, and a few about the brand.
- It collects the answers. Claude runs the panel with its own web search. You run ChatGPT, Perplexity or Gemini by hand and paste the raw answers.
- It scores and diagnoses. score.py computes the score, the competitor ranking and the cited sources. Then come the top three gaps and their cause.
What you get back
- A score out of 100 with its band, its plausible range and its components
- A table per question: named or not, who is named instead, sources
- The sites the engines cite, and whether you are present on each
- An action plan ordered by evidence, and a re-test plan
Run on this very site
A first reading of whether this pack is named, on one web search tool (not an assistant), one run.
Presence is 0 of 2 non-brand runs. [...] No score: it would be indicative only.
"I did not find a product specifically called SEO Skills by PENRA"
321 GitHub repos with "seo-skills" in the name
OAI-SearchBot: allowed | Claude-SearchBot: allowed | PerplexityBot: allowedExpected for a domain two days old: this is the starting point, dated. The name is always written with its publisher, and the same questions will be asked again in four to six weeks.
What it will not do
- On its own, Claude tests one engine: its own web search. It never presents that as ChatGPT, Perplexity or Gemini.
- Answers change from run to run, so the score is a sample. No engine publishes its ranking factors, and the skill promises no citations.
- It will not rig the test with leading questions, or recommend fake reviews, hidden text for AI or paid "AI mention" packages.
What it needs
Claude Code or the Claude apps, with a web search tool for the part Claude runs. The two scripts need Python 3.9 or later and no network. Without them you get presence counts, but no composite score.
How to install a skill in Claude
Guides
Where it hands over
- Street Cred: It turns the cited sources where you are missing into a plan of listings and pitches.
- Schema Sheriff: It produces and checks the schema markup and llms.txt that this audit only points to.
Questions about this skill
Does it query ChatGPT, Perplexity and Gemini itself?
No. Claude runs only its own web search. For the others you paste their raw answers. An engine you skip is reported as not tested.
How much work is the manual test?
A panel is 15 to 30 questions per market. In full mode you ask each one three times per assistant, in a fresh chat.
It comes with the 13 others.
One pack, one payment: the 14 skills, their scripts, templates and reference notes.
One payment, no subscription. In your inbox within a minute, as a rule.