doc-prd
doc-prd turns a conversation you have already had into a published PRD. After an
interview or a doc-align grilling, the shared understanding is sitting in the
context — this skill writes it up and pushes it to the issue tracker, without
asking you everything over again.
What it solves
The gap between “we agreed on this in the discussion” and “there’s a ticket
someone can pick up” is where good thinking evaporates. Re-interviewing to write
the PRD wastes the work already done and risks drifting from what was decided.
And PRDs that leak file paths and code turn into stale implementation specs the
moment the code moves. doc-prd closes both gaps: it captures the decision while
it’s fresh, and keeps the document at the level of intent.
How it works
- No re-interview — the knowledge is assumed to be in context from the prior brainstorm or alignment, so it synthesises rather than re-elicits.
- Intent, not implementation — file paths and code snippets are forbidden in the body, with one exception: a prototype snippet that encodes a decision more precisely than prose would. A PRD describes what and why, not which file.
- Straight to ready-for-agent — it publishes to the issue tracker already labelled for pickup, skipping the triage round-trip.
It is part of the doc-* family: interview and doc-align produce the
understanding, doc-prd publishes it.
When to use
Right after an interview or doc-align session, when the decision is clear and
you want a ticket without re-walking the conversation. If the understanding
isn’t settled yet, run the intake skill first.