What happens when you replace your single most-trusted pitch reviewer with a crowd of 150 strangers who all have opinions?
A developer built synth-personas, a Claude Code skill that does exactly that. Point it at a pitch deck, strategic memo, or any markdown document, and it routes the content through 150 distinct simulated tech personas - covering VCs, skeptical CTOs, first-time customers, and burned-by-SaaS operators, among others. Each persona evaluates the document and returns structured feedback from their own perspective.
The creator built it to solve a specific problem: feedback from a founder friend comes filtered through politeness, one professional background, and the mood of a single afternoon. Running the same document past 150 different readers removes that filter. You stop getting "this looks solid" and start getting "three separate personas flagged your market sizing as unsubstantiated."
Claude Code skills are reusable task definitions - structured prompts with logic - that Claude executes against your project files. The synth-personas skill reads a markdown file, constructs detailed persona profiles, and runs evaluation passes across all of them. A full batch takes roughly 10-15 minutes. Token costs run a few dollars per run depending on document length and which Claude API tier you're using.
The feedback format surfaces patterns rather than directives. No single persona opinion is authoritative, but when 40 out of 150 react the same way to your pricing section, that's a signal worth investigating before you're in a room with actual investors.
The skill isn't limited to fundraising. Product specs, strategic one-pagers, and internal memos work the same way. The broad persona range tends to surface objections that a single reviewer misses because they approach the document from only one angle.
Running it requires Claude Code set up with API access. The persona-simulation approach has existed as a prompting technique for years. What this packages up is the repetition and structure needed to make it practical - the difference between a technique you know works in theory and one you'll actually run before a critical meeting.