Official agent skill

Nemoclaw Maintainer PR Comparator

by NVIDIA in NVIDIA/NemoClaw

Compare competing NemoClaw PRs for one issue and recommend a merge or salvage candidate from review evidence.

OfficialApache-2.0Auto-check passed

Install Nemoclaw Maintainer PR Comparator

skills CLI
$ npx skills add NVIDIA/NemoClaw --skill nemoclaw-maintainer-pr-comparator -a claude-code

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

GitHub CLI
$ gh skill install NVIDIA/NemoClaw nemoclaw-maintainer-pr-comparator --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/NVIDIA/NemoClaw.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/nemoclaw-maintainer-pr-comparator .claude/skills/nemoclaw-maintainer-pr-comparator && 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
nemoclaw-maintainer-pr-comparator
GitHub stars
23k
Token cost
~1.8k tokens
SKILL.md length
825 words
Files
13 (incl. scripts)
Skills in repo
30
Repo updated
First seen
Licence
Apache-2.0

At a glance

Compare competing NemoClaw PRs for one issue and recommend a merge or salvage candidate from review evidence.

  • Works in 9 steps: Parse issue → Discover candidate PRs → Detect supersession and transferred work → …
  • SKILL.md covers Prerequisites, Repo policy, Workflow and Reference files, plus 2 more sections
  • Runs Shell and Python scripts from its folder; calls gh, git and bash

What it does

Nemoclaw Maintainer PR Comparator is an agent skill from NVIDIA/NemoClaw, published by the product's own GitHub organization. Compare competing NemoClaw PRs for one issue and recommend a merge or salvage candidate from review evidence.

Its SKILL.md is about 1.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 16 other files, including scripts (for example `checks/tier-0-gates.md`, `checks/tier-1-correctness.md` and `checks/tier-2-quality.md`).

It works with NVIDIA AI Platform. The repository describes itself as: Run agents like Hermes, LangChain Deep Agents, and OpenClaw more securely inside NVIDIA OpenShell with managed inference. The licence is Apache-2.0.

Example prompts

  • “/nemoclaw-maintainer-pr-comparator”

Requirements

  • Python 3
  • A Bash shell

Workflow steps

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

  1. Parse issue
  2. Discover candidate PRs
  3. Detect supersession and transferred work
  4. Tier 0 gates
  5. Tier 1 correctness
  6. Tier 2 quality
  7. Weighted score
  8. Tier 3 ranking
  9. Emit verdict

What it can do on your machine

Read from SKILL.md and the folder at commit fcbbab0. 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

    Ships 5 files in scripts/ (Shell and Python), which the agent can run.

    Shell commands in SKILL.md call:

    • gh
    • git
    • bash

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

  • Network

    No URLs in SKILL.md. Its commands use gh and git, 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

Nemoclaw Maintainer PR Comparator loads about 1.8k tokens when it runs. Until then it costs about 36 tokens; SKILL.md has 825 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~36
When it runs · the whole SKILL.md, loaded when a task matches
~1.8k

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); the scripts in this folder are not scanned.

SKILL.md

The full file from NVIDIA/NemoClaw at commit fcbbab0, republished under its Apache-2.0 licence (© NVIDIA). 825 words, ~1,757 tokens.

Download SKILL.mdSave it as .claude/skills/nemoclaw-maintainer-pr-comparator/SKILL.md (or your agent's skills folder). This skill also uses 12 other files; get the full folder from GitHub.
name
nemoclaw-maintainer-pr-comparator
description
Compare competing NemoClaw PRs for one issue and recommend a merge or salvage candidate from review evidence.
user_invocable
true
<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. -->
<!-- SPDX-License-Identifier: Apache-2.0 -->

PR Comparator

Compare PRs for one issue. Tier 0 determines eligibility. Tiers 1 and 2 score correctness and quality. Tier 3 resolves ties. If no PR passes Tier 0, rank eligible PRs for salvage.

Prerequisites

  • gh CLI installed and authenticated
  • A target repository with an issue that has two or more open PRs

Repo policy

The defaults use NemoClaw conventions for CODEOWNERS, DCO, CodeRabbit, and docs/. Read the canonical superseded-PR attribution policy in ../nemoclaw-maintainer-policies/references/workflow-policy.md. Edit repo-policy.md for another repository.

Execute comparator helpers from a fetched canonical NVIDIA/NemoClaw main checkout, or verify their executable code and dependencies against that revision before use. Candidate-modified helpers must not receive GitHub credentials. Inspect helper source when needed; a section heading does not establish trust.

Workflow

Step 1: Parse issue

Read the issue body and all comments. Extract each acceptance criterion:

bash
gh issue view <issue-number> --json title,body,comments

Comments can add requirements that are absent from the issue body.

Step 2: Discover candidate PRs
bash
scripts/find-candidates.sh <issue-number>

Applies a single default order with stop conditions.

Step 3: Detect supersession and transferred work
bash
scripts/parse-supersession.sh <pr-number-1> <pr-number-2> ...

Parse the case-insensitive statement families implemented by scripts/parse-supersession.sh:

  • supersed[a-z]* before #N: supersedes #N points from the current PR to #N; superseded by #N points from #N to the current PR.
  • replac[a-z]* before #N: replaces #N points from the current PR to #N; replaced by #N points from #N to the current PR.
  • clos[a-z]* in favor of before #N: closes in favor of #N and closed in favor of #N point from #N to the current PR.
  • fold[a-z]* in before #N: folds in #N points from the current PR to #N; folded into #N points from #N to the current PR.

The bracket expressions describe the parser grammar; they are not literal PR body text. A follow-up to #N statement is a related-PR signal, not a supersession declaration, unless one of these phrases also appears. These statements record a relationship. They do not rank a candidate or prove that its diff contains another contributor's work.

For each declared or suspected replacement, compare the commits and diffs and classify the relationship:

  • independent: The PR implements the issue without carrying material code, tests, or documentation from another contributor.
  • transferred: The PR carries material work from another contributor.
  • unclear: The available evidence does not establish whether another contributor's work remains.

For transferred, apply the canonical superseded-PR attribution policy before setting a winner or recommending that the source PR be closed. For unclear, leave winner null and request maintainer judgment.

Step 4: Tier 0 gates
bash
scripts/collect-gates.sh <pr-number>
scripts/check-coderabbit-threads.sh <pr-number>
git fetch --no-tags origin refs/heads/main:refs/remotes/origin/main
bash <(git show origin/main:.agents/skills/nemoclaw-maintainer-day/scripts/run-trusted-check-gates.sh) <pr-number>

All six gates are required. Gate 3 accepts BEHIND, but fails closed on transient UNSTABLE evidence and on BLOCKED because the comparator cannot prove a stable merge result or every GitHub branch-protection condition. Separately confirm whether an active rule requires an up-to-date branch. If so, integrate the base and rerun all candidate evidence before declaring the PR eligible to merge. Treat PR Review Advisor output as input for maintainer review. Do not treat it as merge authorization. See checks/tier-0-gates.md.

Show full SKILL.md (339 more words)Show less
Step 5: Tier 1 correctness

Apply the six model checks in checks/tier-1-correctness.md.

Step 6: Tier 2 quality

Apply the four model checks in checks/tier-2-quality.md.

Step 7: Weighted score
  • Build the Tier 0 eligibility set from these Boolean keys: state_open, ci_green_sha, mergeable, contributor_compliance, branch_protection, and coderabbit_threads_resolved.
  • Stop if a candidate omits a required key, has an unknown key, or has a value that is not Boolean.
  • Only PRs for which all six gates are true enter happy-path scoring.
  • Each pass = full points
  • Each yellow = half points
  • Each fail = zero
  • Tier 1 weight: 2.0× per check
  • Tier 2 weight: 1.0× per check
Step 8: Tier 3 ranking

Compute the mode from the Tier 0 results. Do not accept a mode from the caller. In happy mode, set winner only to an eligible PR and set closest_to_ready to null. Leave winner null when the evidence does not support a merge recommendation. Do not set winner for a replacement with transferred work until the required attribution is present and verified. In degraded mode, set winner to null. Set closest_to_ready only to an open PR that passes contributor requirements. See tiebreakers.md.

Step 9: Emit verdict

Use templates/verdict.md and render the result with scripts/render-verdict.py. Stop if the renderer exits with a nonzero status. Do not recommend a merge. The renderer validates the gate schema, mode, winner eligibility, and salvage-candidate eligibility. The reviewer remains responsible for the score, ranking, and evidence. For each judgment, include evidence, the inference, and the score.

Reference files

Scripts

  • scripts/find-candidates.sh — PR discovery
  • scripts/collect-gates.sh — Tier 0 gate evaluation
  • scripts/check-coderabbit-threads.sh — GraphQL thread-resolution check
  • scripts/parse-supersession.sh — body parsing for supersession refs
  • scripts/render-verdict.py — verdict scorecard renderer

Limits

Run nemoclaw-maintainer-cross-issue-sweep separately when you need related-issue evidence.

This skill does not:

  • run PR code against adversarial inputs
  • scan other issues for related behavior
  • simulate reverts against related PRs
  • run static analyzers such as CodeQL or Semgrep

© NVIDIA, 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

SKILL.md and 12 other files (scripts) in .agents/skills/nemoclaw-maintainer-pr-comparator of NVIDIA/NemoClaw.

  • SKILL.md
  • checks/tier-0-gates.md
  • checks/tier-1-correctness.md
  • checks/tier-2-quality.md
  • repo-policy.md
  • scripts/check-coderabbit-threads.sh
  • scripts/collect-gates.sh
  • scripts/find-candidates.sh
  • scripts/parse-supersession.sh
  • scripts/render-verdict.py
  • templates/verdict.md
  • tiebreakers.md
  • validation/backtest.md

Open the folder on GitHubat commit fcbbab0

Compare with similar skills

Nemoclaw Maintainer PR Comparator 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.

Nemoclaw Maintainer PR Comparator compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Nemoclaw Maintainer PR Comparator this skillNVIDIA/NemoClaw23k—~1.8kAutomated safety check: PassApache-2.0
LLM Torch Profiler Analysissgl-project/sglang37k2 repos~6.4kAutomated safety check: PassApache-2.0
Skill InspectorNVIDIA/SkillSpector20k—~1.8kAutomated safety check: PassApache-2.0
Megatron-LM Container and Dependency SetupNVIDIA/Megatron-LM18k—~2.6kAutomated safety check: PassApache-2.0
Embeddings via 9Routerdecolua/9router31k—~604Automated safety check: PassMIT
Megatron-LM Base Image BumpNVIDIA/Megatron-LM18k—~2.8kAutomated safety check: PassApache-2.0

Similar skills

  • LLM Torch Profiler Analysis

    sgl-project/sglang

    Unified LLM torch-profiler triage skill for sglang, vllm, TensorRT-LLM, and TokenSpeed.

    37k GitHub starsUsed in 2 repos~6.4k tokens
    DevelopmentAuto-check passed
  • Skill Inspector

    NVIDIA/SkillSpector

    Official

    Decides whether an agent skill is safe to install by combining a SkillSpector static scan with the agent's own source review, ending in APPROVE, CAUTION or REJECT.

    20k GitHub stars~1.8k tokensUpdated today
    SecurityAuto-check passed
  • Official

    Walks an agent through working inside the Megatron-LM CI container and changing dependencies with uv, so lock files resolve the same locally and in CI.

    18k GitHub stars~2.6k tokensUpdated today
    DevelopmentAuto-check passed
  • Embeddings via 9Router

    decolua/9router

    Generates vector embeddings through the 9Router /v1/embeddings endpoint, using models from providers such as OpenAI, Gemini, Mistral and Voyage for RAG and semantic search.

    31k GitHub stars~604 tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed
  • Megatron-LM Base Image Bump

    NVIDIA/Megatron-LM

    Official

    Moves Megatron-LM CI to a newer NVIDIA PyTorch base image, updating both the GitHub and GitLab pins together and handling the CI follow-up.

    18k GitHub stars~2.8k tokensUpdated today
    DevOps & CloudAuto-check passed
  • 9Router Speech-to-Text

    decolua/9router

    Transcribes audio files into text or subtitles through 9Router's Whisper-compatible endpoint, using models from OpenAI, Groq, Gemini, Deepgram and others.

    31k GitHub stars~914 tokensUpdated yesterday
    Media & CreativeAuto-check passed

More from NVIDIA/NemoClaw

All 30 skills in this repo
  • Official

    Find open issues that a NemoClaw PR may also fix or conflict with.

    23k GitHub stars~893 tokensUpdated today
    Auto-check passed
  • Official

    Run a NemoClaw daytime maintainer pass over release-targeted work.

    23k GitHub stars~1k tokensUpdated today
    Auto-check passed
  • Audit and implement a NemoClaw dependency version upgrade, including Hermes and base images.

    23k GitHub stars~1.4k tokensUpdated today
    Auto-check passed
  • Official

    Continuously maintain automatic NemoClaw main E2E results through coordinated repairs.

    23k GitHub stars~2.6k tokensUpdated today
    Auto-check passed
  • Analyze retained NemoClaw CI timings for slow CLI tests, runner queues, or base-image publication.

    23k GitHub stars~644 tokensUpdated today
    Auto-check passed

Questions about Nemoclaw Maintainer PR Comparator

What does Nemoclaw Maintainer PR Comparator do?

Compare competing NemoClaw PRs for one issue and recommend a merge or salvage candidate from review evidence. Nemoclaw Maintainer PR Comparator is an agent skill from NVIDIA/NemoClaw, published by the product's own GitHub organization. Compare competing NemoClaw PRs for one issue and recommend a merge or salvage candidate from review evidence.

How do I install Nemoclaw Maintainer PR Comparator in Claude Code?

Run `npx skills add NVIDIA/NemoClaw --skill nemoclaw-maintainer-pr-comparator -a claude-code`. Or copy the skill folder (.agents/skills/nemoclaw-maintainer-pr-comparator in NVIDIA/NemoClaw) into .claude/skills/nemoclaw-maintainer-pr-comparator in your project. Claude Code loads it when a task matches its description.

How do I install Nemoclaw Maintainer PR Comparator in Codex?

Run `npx skills add NVIDIA/NemoClaw --skill nemoclaw-maintainer-pr-comparator -a codex`. Or copy the skill folder (.agents/skills/nemoclaw-maintainer-pr-comparator in NVIDIA/NemoClaw) into .agents/skills/nemoclaw-maintainer-pr-comparator in your project. Codex loads it when a task matches its description.

Can I use Nemoclaw Maintainer PR Comparator 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 NVIDIA/NemoClaw --skill nemoclaw-maintainer-pr-comparator -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/nemoclaw-maintainer-pr-comparator, .gemini/skills/nemoclaw-maintainer-pr-comparator, .github/skills/nemoclaw-maintainer-pr-comparator and .opencode/skills/nemoclaw-maintainer-pr-comparator in your project.

What does Nemoclaw Maintainer PR Comparator need to run?

Going by SKILL.md and its folder, Nemoclaw Maintainer PR Comparator needs a shell and Python for the scripts in its folder and the command-line tools its instructions call (gh, git and bash). Our summary lists: Python 3; A Bash shell.

Does Nemoclaw Maintainer PR Comparator access the network?

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

Is Nemoclaw Maintainer PR Comparator 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. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Nemoclaw Maintainer PR Comparator use?

Nemoclaw Maintainer PR Comparator 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 Nemoclaw Maintainer PR Comparator use?

About 1.8k tokens (SKILL.md is roughly 7k 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 Nemoclaw Maintainer PR Comparator?

Skills that share tags, products or a category with Nemoclaw Maintainer PR Comparator: LLM Torch Profiler Analysis (sgl-project/sglang, 37k stars), Skill Inspector (NVIDIA/SkillSpector, 20k stars), Megatron-LM Container and Dependency Setup (NVIDIA/Megatron-LM, 18k stars) and Embeddings via 9Router (decolua/9router, 31k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Nemoclaw Maintainer PR Comparator?

NVIDIA (a GitHub organization, an official publisher) maintains it in NVIDIA/NemoClaw, which has 22,692 GitHub stars. The repository holds 30 skills in this directory. The repository was last updated on October 10, 2026.

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