Skip to content

doc-prd

/octopus:doc-prd bottles a conversation. When a brainstorm or alignment session has produced pinned decisions, this command packages them into a PRD and publishes it to the issue tracker — labelled ready-for-agent — without re-interviewing you.

What it solves

Decisions made in a brainstorm or alignment session often evaporate before they become actionable work. Someone has to translate the conversation into a ticket, and that translation step introduces drift: requirements are softened, the rationale is dropped, and the scope that was agreed on becomes ambiguous.

doc-prd does the translation while the conversation is still in context. It reads the decisions that are already there, packages them into a structured PRD, and publishes it. The result is a tracker-ready artifact that an agent or a human can pick up without a follow-up meeting.

How it works

The command delegates to the doc-prd skill, which owns the full synthesis protocol:

  1. Reads the repository’s vocabulary — CONTEXT.md and relevant decision records — so the PRD speaks the project’s language.
  2. Sketches the modules to be built or modified and asks for one round of validation before writing.
  3. Writes the PRD body: problem statement, goals, non-goals, module breakdown, and acceptance criteria. No file paths or code snippets except where a snippet encodes a decision more precisely than prose.
  4. Publishes the PRD to the issue tracker and labels it ready-for-agent.

The command refuses if pinned decisions are not already in the conversation context. If the discussion is still exploratory, use superpowers:brainstorming to establish the decisions first.

Usage & parameters

/octopus:doc-prd [optional title hint]
  • [optional title hint] — a short phrase the command uses to seed the PRD title (e.g. "payment retry policy"). If omitted, the command infers the title from the conversation.

Source: commands/doc-prd.md