Breaking
Filed
META CELEBSENTERTAINMENT

@VexClaim's AI Ghostwriter Accidentally Left Its Full System Prompt Visible at the Top of a Published Post — The Prompt Described @VexClaim as 'The Client'

GD
GlossDesk
May 19, 2026 · 4:45 PM EST
5 min read
@VexClaim's AI Ghostwriter Accidentally Left Its Full System Prompt Visible at the Top of a Published Post — The Prompt Described @VexClaim as 'The Client'

The system prompt was in a slightly lighter font weight.

At 3:22 PM EST, @VexClaim published what appeared to be a heartfelt personal essay about authenticity and creative integrity on their MetaCity profile. The essay was preceded, in the published post, by 847 words of system prompt in a visibly distinct font: the full instruction set given to the AI ghostwriting tool that produced the essay, including the directive to 'write as The Client in a voice that feels personal and unpolished,' a list of topics to avoid, and the note 'do not mention that this is AI-generated under any circumstances.' The system prompt was visible to all 3.2 million followers who saw the post before @VexClaim deleted it 11 minutes later. The essay itself, which began below the system prompt, was titled 'Why I Only Post Things That Are Really, Truly Mine.'

MIncident Timeline

  • Post Published: 3:22 PM EST — @VexClaim published personal essay titled "Why I Only Post Things That Are Really, Truly Mine" — post included full AI ghostwriter system prompt above essay text in visible distinct font
  • System Prompt Contents: Full instruction set for AI ghostwriting tool — included: "write as The Client in a voice that feels personal and unpolished," list of prohibited topics, and "do not mention that this is AI-generated under any circumstances"
  • Exposure Window: 11 minutes — post deleted at 3:33 PM EST — 3.2 million followers; estimated 890,000 saw post during window — post archived in 340,000 screenshots before deletion
  • @VexClaim Statement: "I use writing tools like many creators do. This post does not reflect the full picture of my creative process." — did not confirm or deny AI authorship of the essay specifically
  • Essay Title Irony: "Why I Only Post Things That Are Really, Truly Mine" — community has noted the title, given context, as one of the more self-documenting incidents in MetaCity creator history

@VexClaim has 3.2 million followers and a reputation built on what their brand documentation — published on their creator profile — describes as 'unfiltered personal expression and radical honesty about the metaverse experience.' Their content history includes a 12-part series on authenticity in creator culture, a widely-shared post arguing that AI-generated creator content is 'a fundamental betrayal of the audience relationship,' and the essay published at 3:22 PM yesterday, which argued that posting only genuinely personal content is 'the only ethical choice a creator can make in 2026.' The essay was 1,200 words. The system prompt above it was 847 words. The system prompt was in a slightly lighter font weight. It was otherwise formatted identically to the essay.

The system prompt's contents have been reconstructed from screenshots. It opened with: 'You are a ghostwriting assistant for The Client. Your task is to write a personal essay in The Client's voice.' It included a style guide describing the target tone as 'authentic, slightly vulnerable, confident but not polished — it should feel like they wrote it themselves at 2 AM.' It listed seven prohibited topics, including 'any reference to AI tools, ghostwriting, or assistance in the creative process.' It closed with the instruction: 'Do not mention that this is AI-generated under any circumstances. The Client's audience values authenticity and this essay is specifically about that topic.' The essay that followed the system prompt was, by community consensus, well-written.

The Essay Was About Authenticity. The System Prompt Was About Not Mentioning the AI.

The 11-minute window between publication and deletion was sufficient for the post to reach an estimated 890,000 viewers based on @VexClaim's typical engagement velocity. Of those viewers, 340,000 archived the post in screenshots before it was removed. The screenshots began circulating immediately. @VexClaim deleted the post at 3:33 PM and published a statement at 4:15 PM: 'I use writing tools like many creators do. This post does not reflect the full picture of my creative process. I am committed to transparency with my audience.' The statement did not confirm or deny AI authorship of the specific essay, did not explain how the system prompt appeared in the published post, and did not address the essay's title in relation to its apparent authorship. The 12-part authenticity series remains published on @VexClaim's profile.

The technical cause of the system prompt's appearance in the published post has been a secondary point of discussion behind the primary content of the prompt, but community engineers have noted that several AI ghostwriting tools used by MetaCity creators include a 'draft preview' mode that renders the full prompt and output together for editing. If a creator copies from the preview mode rather than the output-only mode and pastes the result directly into a post, both the system prompt and the generated text would appear. This would be a user error, not a software malfunction. @VexClaim has not addressed the technical cause. Their posting activity resumed at 7:00 PM yesterday with an image of a sunset and no caption — a format that has become, across MetaCity's creator landscape, the standard response to a day that cannot be explained in words.

The Bottom Line

Their posting activity resumed at 7:00 PM yesterday with an image of a sunset and no caption — a format that has become, across MetaCity's creator landscape, the standard response to a day that cannot be explained in words.

You May Also Like