us-ac-formatter
The user gives you user stories and acceptance criteria as an ASCII or markdown table. Convert them into the project's structured output: US-XX story entries, AC-XX.YY Gherkin scenarios, and a sprint breakdown, matching the exact shape in references/layout.md.
Steps:
-
Parse the table. Identify each user story (persona, action/goal, business value) and each acceptance criterion attached to it. One table row is usually one US or one AC; infer the columns from the header. If the table is ambiguous (columns unclear, AC not mapped to a US, no sprint hints), ask before guessing.
-
Assign IDs. Number stories
US-01,US-02, and so on, and their criteriaAC-<story>.<seq>(AC-01.01). Preserve any IDs already present in the input rather than renumbering. If the input has gaps in numbering, keep them. The index script adds the gap notes. -
Group by sprint goal. Every US needs a sprint and every sprint a goal title. If the input carries a sprint breakdown (for example from
product-discovery), confirm it with the user. If it has none, or it is ambiguous, ask. Never invent one. -
Write Gherkin. Convert each AC into a
Given / When / Thenbody inside a ```gherkin fence. UseAndfor continuation lines, indented two spaces. Where a Then enumerates fields/buttons/options, list them as- itembullets. Keep the user's domain wording and language verbatim (e.g. Bahasa Indonesia UI copy): translate nothing. -
Emit the three artifacts in the output format of references/layout.md, writing them into
docs/business/(see its file targets). Read each target first if it exists; merge rather than clobber unless the user asks for a clean overwrite. Show a short summary of what was written.