Prompt sets and leading questions
A product file
Section titled “A product file”A product file names the product, the places that are its own, and the facts the answers should get right:
[product]name = "Modelado"aliases = []places = ["alegauss.com/modelado", "github.com/alegauss/modelado"]phrases = [] # wording distinctive of its own pages, which a neutral question avoids
[[fact]]attribute = "Runs without an NVIDIA GPU"value = "Yes: any DirectX 12 GPU through DirectML, or the CPU"page = "https://alegauss.com/modelado/" # where it should be said plainlyaskround ask "..." --product-file modelado.toml --extract compares what the answers claim
with the facts. A claim that contradicts a fact, or doubts it, is reported as a fact gap
with the page that should state the fact plainly.
A prompt set
Section titled “A prompt set”Add [[question]] entries and the file becomes the set of questions you track:
[[question]]text = "What app can turn a photo into a 3D model on Windows without an NVIDIA GPU?"purpose = "constraint" # generic, constraint, use case, languagelocale = "BR:pt-BR" # optionalruns = 3 # optionalaskround run modelado.toml (or askround_run over MCP) asks every question as its own
round, kept together in one set directory. Run it every week from the system’s scheduler,
with --max-cost so an unattended run spends no more than you agreed to.
The MCP server lists the product files it finds at the top of the project and one
directory down, with their question counts, in askround_run’s schema.
Questions askround refuses
Section titled “Questions askround refuses”A question that names the product it tests (its name, an alias, one of its places, or a
phrase under phrases) measures nothing: the answer names the product because the question
did. askround refuses it before anything is paid for. A question that asks about the
product on purpose is marked brand = true (or --brand); it runs and is reported apart.
Ask the way the product’s audience asks: their constraint (“without an NVIDIA GPU”), their
use case, their language, with locale set and the question written in it.
Which wording leads to the product?
Section titled “Which wording leads to the product?”askround discover modelado.toml --seed "free offline app photo to 3D model on Windows"From the seed and the product’s facts, a model writes variants (a constraint the product meets, a use case, another language) that never name it. askround runs them with the seed and ranks them by how often and how high the product appears. One run per question is a sketch: put the promising ones in the product file and run them with more.