pr-summary
- Repo stars 2
- License Apache-2.0
- Author updated Live
- Author repo ai-dev
- Domain
- Engineering
- Compatible agents
-
- Claude Code
- Cursor
- Cline
- Codex
- Windsurf
- Gemini CLI
- +20
- Trust score
- 94 / 100 · audit passed
- Author / version / license
- @LLM-Dev-Studio · Apache-2.0
- Token usage
- Lean
- Setup complexity
- Plug-and-play
- External API key
- Not required
- Operating systems
- Unspecified (assume cross-platform)
- Runtime requirements
- No special requirements
- Permissions
-
- Read-only
- Write / modify
- Network behavior
- Local-only
- Install commands
- 26 variants
Profile is derived at build time from SKILL.md and install vectors. Subject to drift from author intent.
Heads up: 未限定 allowed-tools,默认拥有全部工具权限。
---
name: pr-summary
description: Generates a structured PR summary for the current branch by inspecting commits and file changes…
category: engineering
runtime: no special runtime
---
# pr-summary output preview
## PART A: Task fit
- Use case: Generates a structured PR summary for the current branch by inspecting commits and file changes since master. Use when creating a PR or when asked for a summary of branch changes..
- Inputs: target material, constraints, expected output, and acceptance criteria.
- Evidence boundary: follow “Workflow / 1. Gather branch context / 2. Analyse the changes” and do not present inference as author intent.
## PART B: Execution result
- **01** The card summarizes the use case; runtime output centers on “Generates a structured PR summary for the current branch by inspecting commits and file changes since master. Use when creating a PR or when asked for a summary of branch changes.”.
- **02** When the source has headings, the agent prioritizes “Workflow / 1. Gather branch context / 2. Analyse the changes” so the result follows the author’s structure.
- **03** Typical output includes task judgment, concrete steps, required commands or file edits, validation, and follow-up options.
- **04** Risk context follows the fingerprint: read files, write/modify files; mostly runs locally; usually needs no extra API key.
## Running Rules
- read files, write/modify files; mostly runs locally; usually needs no extra API key.
- Validate with a small sample before expanding scope.
- Return the result, validation criteria, and next iteration options. The source mentions slash commands such as `/code-review`; use them first when your agent supports command triggers.
Name target files or source material, expected output, forbidden changes, and whether network or shell access is allowed. Permission fingerprint: read files, write/modify files.
Start with a small task and check whether the result follows “Workflow / 1. Gather branch context / 2. Analyse the changes”. Inspect diffs, logs, previews, or tests before expanding scope.
Confirm the final output includes a concrete result, evidence, and next action. If it stays generic, tighten inputs, boundaries, and acceptance criteria.
---
name: pr-summary
description: Generates a structured PR summary for the current branch by inspecting commits and file changes…
category: engineering
source: LLM-Dev-Studio/ai-dev
---
# pr-summary
## When to use
- Generates a structured PR summary for the current branch by inspecting commits and file changes since master. Use when…
- Use it when the task has clear inputs, repeatable steps, and validation criteria.
## What to provide
- Target material, scope, expected result, and forbidden changes.
- Whether network, commands, file writes, or external services are allowed.
## Execution rules
- Organize steps around “Workflow / 1. Gather branch context / 2. Analyse the changes” and keep inference separate from source facts.
- read files, write/modify files; mostly runs locally; usually needs no extra API key.
- Validate with a small sample before expanding the task.
## Output requirements
- Return the deliverable, key evidence, validation method, and next action.
- Mark missing information as unknown; do not invent commands, platforms, or dependencies. The author source anchors workflow facts; repository files anchor sources and commands; Fluxly only adds fit, limitations, and quality judgment.
skill "pr-summary" {
input -> user goal + target files + boundaries + acceptance criteria
context -> Workflow / 1. Gather branch context / 2. Analyse the changes
rules -> SKILL.md triggers / order / output contract
runtime -> no special runtime | read files, write/modify files | mostly runs locally
guardrails -> usually needs no extra API key + small-sample validation + diff/log review
output -> copyable result + checklist + next iteration
} PR Summary
This skill generates a structured summary of all changes on the current branch relative to a base branch (default: master).
Workflow
1. Gather branch context
git branch --show-current
git log master..HEAD --oneline
git diff master...HEAD --stat
git log master..HEAD --format="%s%n%b"
If the user supplied a different base branch as an argument, substitute it for master.
Extract the work item number from the branch name (first run of 4+ digits):
| Branch | Work item |
|---|---|
feature/12345-add-foo |
12345 |
feedback/58340/from-luke |
58340 |
JTD-56566-missing-screen |
56566 |
docs/59478 |
59478 |
2. Analyse the changes
Group changes into logical themes by reading the commit messages and diff stat. Common themes for this project:
- New features — new screens, controls, device integrations, service bus consumers
- Bug fixes — correctness fixes (mention what was wrong and what was corrected)
- Renames / breaking changes — entity or contract renames, API changes
- Service bus / messaging — MassTransit consumers, message contracts, RabbitMQ configuration
- Device integration — RFID, camera, BlackMoth, or other device changes
- UI changes — WPF views, XAML resources, InTruck or Console screens
- Dead code removal — deleted files, unused services, obsolete utilities
- Tooling / config — CLAUDE.md, skills, CI, migrations, feature flags
3. Produce the title and description
Title — one line following the project's PR title format: JTD <work-item> - Brief description
- Use the work item number extracted from the branch name
- Brief description: imperative, ≤ 70 characters total
- Example:
JTD 58106 - Handle missing RORO bin weights using tipping records
Output the title first in a code block, then the description below it.
Description — output the entire description inside a single fenced markdown code block (```markdown) so the user can copy-paste it directly into Azure DevOps. Use this structure inside the block:
```markdown
## Description — Work Item #<id>
### <Theme 1 — e.g. "New Feature: Trailer Management Screen">
- Bullet describing the change. Mention affected components (InTruck, Console, ServiceBus, Shared, Devices) where relevant.
### <Theme 2>
- ...
### Bug Fixes
- **<Short label>** — what was wrong and what was changed to fix it.
### Dead Code / Cleanup
- List deleted files or removed abstractions with a one-line rationale.
### Tooling & Config
- Any non-code changes (CLAUDE.md, skills, CI, feature flags).
---
**Files changed:** X | **Insertions:** +Y | **Deletions:** -Z
```
Rules:
- Use
###headings for each theme; omit sections that have no relevant changes. - Keep bullets concise — one sentence per item, no padding.
- Always include the
Work Item #<id>in the description heading. - Always include the files/insertions/deletions line at the bottom (read from
git diff --statsummary line). - If the branch has zero commits ahead of master, say so and exit.
- The entire description (everything inside the markdown code block) must not exceed 3500 characters. If the draft exceeds this limit, shorten bullets — summarise groups of related changes into a single bullet, drop minor cleanup items, and abbreviate file lists — until it fits. Never truncate mid-sentence; always produce complete, grammatical bullets.
4. Offer next steps
After the summary, ask if the user wants to:
- Create the PR now (
gh pr create) using the summary as the body - Review the changes first (
/code-review) - Adjust any section of the summary
Decide Fit First
Design Intent
How To Use It
Boundaries And Review