New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Official agent skill

Review PR

by intel in intel/auto-round

Review or prepare a pull request for the AutoRound repository — checks registration points for new data types/backends/VLMs, validates Chinese translation parity for modified markdown files…

OfficialApache-2.0Auto-check passedDevelopment

Install Review PR

skills CLI
$ npx skills add intel/auto-round --skill review-pr -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install intel/auto-round review-pr --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/intel/auto-round.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/review-pr .claude/skills/review-pr && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
review-pr
GitHub stars
1.6k
Token cost
~1.7k tokens
SKILL.md length
684 words
Files
1
Skills in repo
7
Repo updated
First seen
Licence
Apache-2.0

At a glance

Review or prepare a pull request for the AutoRound repository — checks registration points for new data types/backends/VLMs, validates Chinese translation parity for modified markdown files…

  • Works in 6 steps: Code Quality → Quantization-Specific Concerns → Registration Points → …
  • Performing a code review
  • SKILL.md covers Review Sequence, Review Checklist, Chinese Translation Verification and Common Issues to Watch For
  • Calls git and pip

What it does

Review PR is an agent skill from intel/auto-round, published by the product's own GitHub organization. Review or prepare a pull request for the AutoRound repository — checks registration points for new data types/backends/VLMs, validates Chinese translation parity for modified markdown files, verifies quantization numerical stability (scale overflow, STE gradient flow, groupsize padding), confirms test placement and fixture usage, and enforces Apache 2.0 headers and DCO sign-off. Use when performing a code review, running a PR checklist, preparing a merge request, or auditing a contribution before submit.

Its SKILL.md is about 1.7k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development, covering Pull requests, LLM inference and serving and Translation. The repository describes itself as: A simple and effective post training quantization toolkit for high-accuracy low-bit LLM inference|简洁且高效的后训练量化工具包. The licence is Apache-2.0.

When your agent uses it

  • Performing a code review
  • Running a PR checklist
  • Preparing a merge request
  • Auditing a contribution before submit

Example prompts

  • “/review-pr”

Requirements

  • Python 3

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. Code Quality
  2. Quantization-Specific Concerns
  3. Registration Points
  4. Test Coverage
  5. Documentation
  6. Contributing Requirements

What it can do on your machine

Read from SKILL.md and the folder at commit ae21ef9. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git
    • pip

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git and pip, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Review PR loads about 1.7k tokens when it runs. Until then it costs about 130 tokens; SKILL.md has 684 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~130
When it runs · the whole SKILL.md, loaded when a task matches
~1.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from intel/auto-round at commit ae21ef9, republished under its Apache-2.0 licence (© intel). 684 words, ~1,721 tokens.

Download SKILL.mdSave it as .claude/skills/review-pr/SKILL.md (or your agent's skills folder).
name
review-pr
description
Review or prepare a pull request for the AutoRound repository — checks registration points for new data types/backends/VLMs, validates Chinese translation parity for modified markdown files, verifies quantization numerical stability (scale overflow, STE gradient flow, group_size padding), confirms test placement and fixture usage, and enforces Apache 2.0 headers and DCO sign-off. Use when performing a code review, running a PR checklist, preparing a merge request, or auditing a contribution before submit.

Pull Request Review Workflow for AutoRound

Review Sequence

Follow these steps in order. Stop and request changes at any gate that fails.

  1. Scope check — read the PR description and diff summary. Confirm the PR does one thing and unrelated changes are absent. If scope is unclear, request changes before proceeding.
  2. Pre-commit validation — install pre-commit if needed and run all configured checks:
bash
pip install pre-commit
pre-commit run --all-files

Any failure → request changes. 3. Code quality gate — run through the Code Quality checklist below. Any failure → request changes. 4. Quantization review — if the PR touches auto_round/ quantization logic, run the Quantization-Specific checklist. Any numerical stability concern → request changes. 5. Registration audit — if the PR adds a new feature type (data type, export format, VLM, backend, dataset, scheme), verify every registration point in the table below is updated. Missing registration → request changes. 6. Test verification — confirm new functionality has tests in the correct backend directory with minimal iterations. Missing or misplaced tests → request changes. 7. Documentation & translation — check README/docs updates and run the Chinese Translation Verification procedure below. Missing _CN.md updates for modified markdown → request changes. 8. Contributing requirements — verify DCO sign-off, clean commits, and clear PR description. 9. Decision — if all gates pass, approve. Otherwise, summarize all findings in a single review comment with specific file:line references.

Review Checklist

1. Code Quality
  • Code follows existing patterns in the codebase (decorator registration, factory patterns, etc.)
  • No hardcoded paths or credentials
  • Proper error handling at system boundaries
  • No unnecessary abstractions or over-engineering
  • Import organization follows existing conventions
  • Apache 2.0 license header present on new files:
    python
    # Copyright (c) 2025 Intel Corporation
#
# Licensed under the Apache License, Version 2.0 (the "License");
# ...
2. Quantization-Specific Concerns
  • Numerical stability: scale computation avoids division by zero
  • Gradient flow: uses round_ste() or equivalent STE for differentiable rounding
  • Tensor shapes: group_size reshaping handles padding correctly
  • dtype consistency: scale_dtype, compute_dtype used properly
  • Memory efficiency: no unnecessary tensor copies on GPU
  • Device handling: tensors moved to correct device before operations
3. Registration Points

When the PR adds new functionality, verify all registration points are updated:

FeatureRegistration Location
Data typeauto_round/data_type/__init__.py import + @register_dtype
Export formatauto_round/formats.py @OutputFormat.register()
VLM modelspecial_model_handler.py SPECIAL_MULTIMODAL_BLOCK + lists
Backendauto_round/inference/backend.py BackendInfos dict
Datasetauto_round/calib_dataset.py @register_dataset
Scheme presetauto_round/schemes.py PRESET_SCHEMES dict
4. Test Coverage
  • New functionality has corresponding tests
  • Tests use existing fixtures (tiny_opt_model_path, dataloader, etc.)
  • Tests are placed in the correct backend directory (test_cpu/, test_cuda/, etc.)
  • Tests use minimal iterations (iters=2, nsamples=2) for speed
  • No flaky assertions (avoid exact float comparisons)
Show full SKILL.md (297 more words)Show less
5. Documentation
  • README.md updated if user-facing features change
  • Chinese translation updated: Any changes to *.md files must have corresponding updates in their *_CN.md counterparts:
    • README.md → README_CN.md
    • docs/step_by_step.md → docs/step_by_step_CN.md
    • docs/environments.md → docs/environments_CN.md
  • Translation maintains equivalent content and structure (not just copied English text)
  • Docstrings added for new public APIs
6. Contributing Requirements
  • Commits are signed off (git commit -s) per DCO
  • No unrelated changes mixed in
  • PR description clearly explains the motivation and changes
  • Breaking changes are called out explicitly

Chinese Translation Verification

This is a hard requirement for the AutoRound project. Use this procedure:

  1. Identify modified markdown files:

    bash
    git diff --name-only HEAD~1 -- '*.md'

  2. Check for corresponding CN files: For each modified .md file, verify a _CN.md counterpart exists and is also modified:

    • README.md → README_CN.md
    • docs/step_by_step.md → docs/step_by_step_CN.md
    • docs/environments.md → docs/environments_CN.md

  3. Compare structure:

    • Same number of sections/headings
    • Same tables, code blocks, and links
    • Equivalent content (not machine-translated gibberish)

  4. Files that do NOT need CN translation (no _CN counterpart exists):

    • CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md
    • test/README.md
    • docs/publication_list.md, docs/tips_and_tricks.md, accuracy result docs

Common Issues to Watch For

Quantization Bugs
  • Scale overflow: Large models with small group_size can produce FP16 overflow in scales. Check for torch.clamp or torch.finfo guards.
  • Asymmetric zero-point drift: Zero-points must be integer-rounded for INT quantization.
  • GGUF super-block alignment: GGUF formats require specific block sizes (typically 256 elements). Verify padding/alignment logic.
Export Compatibility
  • Format detection: Verify quantize_config.json or equivalent metadata is saved correctly for the target framework to detect.
  • Weight name mapping: Ensure packed weight names match what the inference framework expects.
  • Mixed-precision layers: Layers excluded from quantization (e.g., lm_head) must be saved in their original format.
Backend Selection
  • Priority conflicts: New backends should not override existing backends unless intentional. Check priority values.
  • Feature checker coverage: Ensure checkers don't silently reject valid layers (test with real model shapes).

© intel, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/review-pr of intel/auto-round.

Open the folder on GitHubat commit ae21ef9

Compare with similar skills

Review PR next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Review PR compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Review PR this skillintel/auto-round1.6k—~1.7kAutomated safety check: PassApache-2.0
Perfupraullenchai/Rapid-MLX3.9k—~1.6kAutomated safety check: NotesCustom licence
Qiaomu Meta Skilljoeseesun/qiaomu-meta-skill383—~2.8kAutomated safety check: PassMIT
Review PRvllm-project/vllm-omni7.1k—~3.8kAutomated safety check: PassApache-2.0
Harness ContributingFairladyZ625/harness-anything225—~4.2kAutomated safety check: PassAGPL-3.0
Review PRjuliepy/AI-Engineer-from-scrach433—~946Automated safety check: NotesNone

Similar skills

  • Perfup

    raullenchai/Rapid-MLX

    Autonomous performance optimization: research, PoC, benchmark, implement, review, PR

    3.9k GitHub stars~1.6k tokensUpdated today
    AI & LLM EngineeringAuto-check: notes

  • Qiaomu Meta Skill

    joeseesun/qiaomu-meta-skill

    Research, create, improve, migrate, evaluate, package, install-check, govern, and safely publish qiaomu-flavored agent skills from workflows, prompts, transcripts, docs, SOPs, runbooks, scripts, or…

    383 GitHub stars~2.8k tokensUpdated 2 mo ago
    DevelopmentAuto-check passed

  • Review PR

    vllm-project/vllm-omni

    Review pull requests and local branches for vllm-project/vllm-omni with a frozen snapshot, module-design ownership, feature-design overlays, targeted validation, and concise evidence-backed findings.

    7.1k GitHub stars~3.8k tokensUpdated today
    AI & LLM EngineeringAuto-check passed

  • Harness Contributing

    FairladyZ625/harness-anything

    Contribute a public change to Harness Anything from a GitHub issue through an isolated worktree, scoped tests, manifest-selected gates, a complete bilingual PR body, review triage, and maintainer…

    225 GitHub stars~4.2k tokensUpdated today
    DevelopmentAuto-check passed

  • Review PR

    juliepy/AI-Engineer-from-scrach

    Review an incoming community PR for waku-agent and present it Sean's way — bilingual (English + 中文), three fixed sections: what they did & why it matters, verdict (merge / change / close), and how…

    433 GitHub stars~946 tokensUpdated 1 mo ago
    DevelopmentAuto-check: notes

  • Docs PR Metadata Guard

    pingcap/docs

    A skill your agent uses when creating or editing pull requests in pingcap/docs so the PR template sections, version checkboxes, related-link fields, HTML comments, and description structure stay…

    616 GitHub stars~1.9k tokensUpdated yesterday
    DevelopmentAuto-check passed

More from intel/auto-round

  • Adapt New Diffusion Model

    intel/auto-round

    Official

    Adapt AutoRound to support a new diffusion model architecture (DiT, UNet, hybrid AR+DiT).

    1.6k GitHub stars~2.8k tokensUpdated today
    Auto-check passed

  • Adapt New LLM

    intel/auto-round

    Official

    Adapt AutoRound to support a new LLM architecture that doesn't work out-of-the-box.

    1.6k GitHub stars~2.4k tokensUpdated today
    Auto-check passed

  • Add Export Format

    intel/auto-round

    Official

    Add a new model export format to AutoRound (e.g., autoround, autogptq, autoawq, gguf, llmcompressor).

    1.6k GitHub stars~1.9k tokensUpdated today
    Auto-check passed

  • Add Inference Backend

    intel/auto-round

    Official

    Add a new hardware inference backend to AutoRound for deploying quantized models (e.g., CUDA/Marlin, Triton, CPU, HPU, ARK).

    1.6k GitHub stars~2.3k tokensUpdated today
    Auto-check passed

  • Add Quantization Datatype

    intel/auto-round

    Official

    Add a new quantization data type to AutoRound (e.g., INT, FP8, MXFP, NVFP, GGUF variants).

    1.6k GitHub stars~1.5k tokensUpdated today
    Auto-check passed

  • Add Vlm Model

    intel/auto-round

    Official

    Add support for a new Vision-Language Model (VLM) to AutoRound, including multimodal block handler, calibration dataset template, and special model handling.

    1.6k GitHub stars~2.4k tokensUpdated today
    Auto-check passed

Categories

Questions about Review PR

What does Review PR do?

Review or prepare a pull request for the AutoRound repository — checks registration points for new data types/backends/VLMs, validates Chinese translation parity for modified markdown files…. Review PR is an agent skill from intel/auto-round, published by the product's own GitHub organization.0 headers and DCO sign-off.

When should I use Review PR?

Review PR fits situations like: performing a code review; running a PR checklist; preparing a merge request; auditing a contribution before submit.

How do I install Review PR in Claude Code?

Run `npx skills add intel/auto-round --skill review-pr -a claude-code`. Or copy the skill folder (.claude/skills/review-pr in intel/auto-round) into .claude/skills/review-pr in your project. Claude Code loads it when a task matches its description.

How do I install Review PR in Codex?

Run `npx skills add intel/auto-round --skill review-pr -a codex`. Or copy the skill folder (.claude/skills/review-pr in intel/auto-round) into .agents/skills/review-pr in your project. Codex loads it when a task matches its description.

Can I use Review PR in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add intel/auto-round --skill review-pr -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/review-pr, .gemini/skills/review-pr, .github/skills/review-pr and .opencode/skills/review-pr in your project.

What does Review PR need to run?

Going by SKILL.md and its folder, Review PR needs the command-line tools its instructions call (git and pip). Our summary lists: Python 3.

Does Review PR access the network?

SKILL.md contains no URLs. Its commands use git and pip, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Review PR safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Review PR use?

Review PR is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Review PR use?

About 1.7k tokens (SKILL.md is roughly 6.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Review PR?

Skills that share tags, products or a category with Review PR: Perfup (raullenchai/Rapid-MLX, 3.9k stars), Qiaomu Meta Skill (joeseesun/qiaomu-meta-skill, 383 stars), Review PR (vllm-project/vllm-omni, 7.1k stars) and Harness Contributing (FairladyZ625/harness-anything, 225 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Review PR?

intel (a GitHub organization, an official publisher) maintains it in intel/auto-round, which has 1,628 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on October 8, 2026.

Source: intel/auto-round on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.