AGENT SKILL

draft-pr

Create a draft PR for the current branch. Use when asked to draft, create, open, or prepare a PR, or to write a PR description.

License 许可
MIT
所属仓库
lithammer/ai
最近更新
2026-11-13
Stars
★ 0
Forks
0

01 前言 Front Matter

name
draft-pr
description
Create a draft PR for the current branch. Use when asked to draft, create, open, or prepare a PR, or to write a PR description.

02 资源文件 3

03 SKILL.md 正文

Draft PR

A fresh subagent writes the description. It reads the branch cold, which is what keeps the body short; the main session supplies only what the repository cannot say.

  1. Identify the current branch and an appropriate base branch.
  2. Write the context: what prompted the change (a ticket, a thread, a conversation, an incident), constraints agreed outside the repository, where the PR sits in a stack, and whether it opens as a draft or ready for review. Facts only, never a body: the drafter writes the prose.
  3. Spawn a subagent. Pass it DRAFTER_BRIEF.md as its persona and DRAFTER_REPORT_FORMAT.md as its output specification, with the branch, the base, and the context as its input.
  4. Relay the URL. Verify each claim the report marks as inferred and correct a wrong one with gh pr edit, that sentence only. If the report says the PR was not created, hand the user its title, body, and command.

04 同仓库更多 Skills

lithammer
lithammer/ai

shape

Review a working change for its shape (cause, ownership, states, names, size, reuse, cost), then apply what you accept.

★ 0
lithammer
lithammer/ai

correctness

Review a working change for defects (logic, state, lifecycle, concurrency, regression) and for drift from its ticket, docs, and rules, then apply what you accept.

★ 0
lithammer
lithammer/ai

zero-tech-debt

Rework a change as if the intended UX and architecture existed from day one, deleting compatibility cruft and accidental complexity.

★ 0
lithammer
lithammer/ai

waterfall-review

Review a code change by dispatching specialized review subagents in parallel — each focused on a single dimension (correctness, security, architecture, etc.) — then triaging their findings into a unified report. Works standalone on a branch or as part of the waterfall flow after implementation. Invoke when the user explicitly asks to review a branch or diff, or as a step inside waterfall-implement; avoid auto-triggering on casual mentions of "review".

★ 0
lithammer
lithammer/ai

waterfall-refactor

Identify refactoring opportunities in code that a feature will touch. Runs at strategic level during design (looking for structural moves that could change which approach is chosen) or tactical level before implementation (looking for pre-work that gives the feature a better foundation). Primarily invoked by waterfall-design or waterfall-implement; avoid auto-triggering on casual mentions of "refactor".

★ 0
lithammer
lithammer/ai

waterfall-plan

Detail a chosen design approach into an implementation plan (objective, file manifest, execution order, test strategy), then auto-proceed to waterfall-challenge. Convergent phase of the waterfall flow.

★ 0