Skip to content
AI & Agents
Skill

/pr-summarizer

Use when opening a PR — produces a clean PR description (what / why / how to verify / risks) from a branch diff against base.

BOOST
From plugin
llm-agents-ecosystem-handbook
55510 skills
Install
$ npx -y skills add oxbshw/LLM-Agents-Ecosystem-Handbook --skill pr-summarizer --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/pr-summarizer

Context preview

The summary Claude sees to decide when to auto-load this skill.

Use when opening a PR — produces a clean PR description (what / why / how to verify / risks) from a branch diff against base.

SKILL.md

pr-summarizer.SKILL.md
name: pr-summarizer
description: Use when opening a PR — produces a clean PR description (what / why / how to verify / risks) from a branch diff against base.
version: 0.1.0
status: experimental
risk: low
tags: [engineering, read-only, writes-files]

PR Summarizer

When to use

  • Just before opening a PR
  • After force-pushing a substantive update to an existing PR (refresh description)

When NOT to use

  • Drafts the user is still iterating on (the summary will go stale immediately)
  • Massive diffs (> 1000 lines) — ask the user to scope first

Inputs

| Name | Type | Required | Notes | |---|---|---|---| | `base` | string | no | default `origin/main` | | `branch` | string | no | default current `HEAD` |

Outputs

A markdown PR body with sections: **Summary**, **Why**, **How tested**, **Risks**, **Linked**.

Workflow

1. `git diff <base>...HEAD --stat` and `git log <base>..HEAD --oneline` to see scope 2. `git diff <base>...HEAD` for content 3. Group changes by area (one bullet per logical concern, not per file) 4. Pull "why" from commit messages + linked tickets if mentioned 5. Identify test artifacts touched; surface as **How tested** (or flag if none) 6. Risks: any High-risk paths touched? auth, billing, migrations? mention. 7. Linked: parse `Fixes #` / `Refs #` from commits

References

  • [`references/pr-template.md`](references/pr-template.md)

Success criteria

  • Description ≤ 25 lines
  • Every "why" claim is supported by a commit message or visible in the diff
  • Risks section honest about unknowns ("untested in production data shape")

Failure modes

  • No commit messages worth quoting → "Why" sourced from diff only; flag this
  • Diff touches secrets / `.env*` → STOP and surface; PR shouldn't include them
Read more
Ships withllm-agents-ecosystem-handbook

A practical operating manual for building, evaluating, securing, and shipping modern LLM agent systems. Modern agents are not "a prompt + a tool."

Get the whole plugin