Questionnaire docs

Get started

Install the skill and start a project-local questionnaire run.

A run starts in a real project directory and creates artifacts under .questionnaire/<timestamp>/.

Install

Install the skill from the public repository, then invoke it from the project you want the artifacts to describe.

install
npx skills add srinitude/questionnaire

Start a run

Run the initializer from a project root. It creates the run directory, copies the standalone HTML, writes the first state.json, and opens index.html in your default browser.

start
python3 scripts/start-questionnaire.py
python3 scripts/start-questionnaire.py --project-dir /path/to/project

Answer the first question

The agent rewrites the starter question for the actual task, presents one question with a recommended answer, and explains why the question matters.

You can answer in chat or edit the local browser draft. The durable project files are still written by the agent.

Keep the run moving

  • After each answer, state.json and transcript.md are updated.
  • CONTEXT.md changes only when a reusable term needs a definition.
  • research/ changes only when an external claim, dependency, or uncertainty needs evidence.
  • adrs/ changes only when the answer creates a hard-to-reverse decision.