What problem does it solve? Auto-generated PR descriptions are usually walls of text that summarize the diff without communicating intent, rationale, tradeoffs, or rejected alternatives. This Skill interviews the PR author to surface the context the code cannot supply, then co-authors an accurate, human-accountable PR description. ## Core Features & Use Cases - Structured author interview: Inspects the PR with git and gh, builds an inventory of design decisions and gaps, then asks one targeted question at a time until author and agent share an accurate understanding. - Co-authored PR body: Writes a What / Why / Changes description with an optional opt-in Requested reviewer focus section, representative examples for user-facing changes, and a provenance footer naming the model and harness. - Consumability report: Verdicts on size, title honesty, narrative, and reviewer focus, plus open topics and unencoded durable policy discovered during the interview. - Use Case: After pushing a feature branch, run the skill from the checked-out branch; it challenges unclear reasoning, lets you amend the PR mid-interview, and publishes an honest description reviewers can act on. ## Quick Start Run /pr-make-accountable from the checked-out branch of an existing PR in a repo with git and an authenticated gh session, optionally adding a sentence of rationale after the invocation.
