Up to eight questions before the spec

Why there is an interview

The spec route is only as good as what it is given. A process flow captures step order, decisions and failure paths, but a build spec also needs actors and permissions, the data behind each step, edge cases, external systems and rough volumes. That knowledge lives with you, so Runic asks before it writes. The interview is stage one of the Build Package; the spec is stage two, and it receives your answers together with the flows.

What it asks

After you click Start, the ticked processes go to the model with Runic's reading legend, so captions, .fail outcomes and .back arrows are read as what they are. The prompt casts the model as a senior business analyst preparing a build specification and asks for five to eight questions, each specific to these flows: it must name the steps or processes it is asking about. Coverage is set by the prompt: actors and roles with permissions, the data each step reads or writes, failure and edge cases, external systems, and volume or constraints.

Questions come back one per line. Runic strips any numbering or bullets the model added and keeps at most eight. The prompt asks for them in the language the flows are written in.

How answers reach the spec

Each question gets a two-line answer box you can drag taller. When you click Generate package, every question is sent with its answer; a blank one is sent as (not answered). The spec prompt has a rule for that case: anything still unknown goes into open_questions instead of being invented. That list renders as the Open Questions section of docs/spec.md, and CLAUDE.md tells Claude Code to consult it and ask you rather than guess.

The full interview is also appended to the spec verbatim, question by question, under Appendix: Interview, so the reasoning behind the spec travels with it.

Limits and gotchas

The questions stage runs on claude-sonnet-5; the spec stage on claude-opus-5. Both count against one limit of 12 calls per hour per account. There is no re-ask button: close the window and click Build again for a fresh set. This is a different interview from the questions that appear in the chat after Create a process with AI returns a linked system; those are bound to a process and turn into edits.

How to use it

1. 1
   Pick the flows

   Click Build in the top bar and tick the processes to include. With nothing ticked, Start refuses with Tick at least one process — the spec is built from what you select.
2. 2
   Start

   Click Start. The window shows Reading your flows… and Finding what a developer would still ask. Esc and the ✕ are inert until the questions arrive.
3. 3
   Read the questions

   They arrive numbered, each with its own answer box. Drag a box's corner to make it taller.
4. 4
   Answer what you can

   The note above the list says it: unanswered questions land in the spec's Open Questions so nothing gets invented.
5. 5
   Generate

   Click Generate package. If the AI call fails, the error shows in an amber note and you land back on the answers with everything you typed intact.

Commands & shortcuts

Tips

• Answer with names and numbers, not adjectives. Finance lead approves refunds over 500 gives the spec a role and a threshold; the right person approves gives it neither.
• Skip honestly. A blank answer is folded into Open Questions, which CLAUDE.md tells Claude Code to consult and ask about before inventing anything.
• Long answers are fine, but the model reads the first 2,000 characters of each. The full text still lands in the spec's appendix.
• If the questions feel generic, the flows are thin. Add the decision or the .fail path in the editor, close the window and click Build again.

Questions