Skip to content

Add Claude Code plugin marketplace for ModelOpt agent skills#1141

Open
kaix-nv wants to merge 1 commit intomainfrom
kaix/agent-marketplace
Open

Add Claude Code plugin marketplace for ModelOpt agent skills#1141
kaix-nv wants to merge 1 commit intomainfrom
kaix/agent-marketplace

Conversation

@kaix-nv
Copy link
Copy Markdown
Contributor

@kaix-nv kaix-nv commented Mar 30, 2026

What does this PR do?

Type of change: ?
New feature. Add Claude Code plugin marketplace manifest so ModelOpt agent skills can be discovered and installed with a single command.

Usage

# Add the ModelOpt marketplace (one time)
/plugin marketplace add NVIDIA/Model-Optimizer

# Install all agent skills (ptq, deployment, evaluation, orchestrator)
/plugin install modelopt-agent@NVIDIA/Model-Optimizer

This bundles all ModelOpt agent skills into one installable plugin:
- /modelopt-agent:ptq — quantize models with FP8, NVFP4, INT4
- /modelopt-agent:deployment — serve checkpoints via vLLM, SGLang, TRT-LLM
- /modelopt-agent:evaluation — benchmark accuracy via NeMo Evaluator Launcher
- /modelopt-agent:modelopt — end-to-end orchestrator (chains PTQ → deploy/eval)

Testing

Before your PR is "Ready for review"

Make sure you read and follow Contributor guidelines and your commits are signed (git commit -s -S).

Make sure you read and follow the Security Best Practices (e.g. avoiding hardcoded trust_remote_code=True, torch.load(..., weights_only=False), pickle, etc.).

  • Is this change backward compatible?: ✅ / ❌ / N/A
  • If you copied code from any other sources or added a new PIP dependency, did you follow guidance in CONTRIBUTING.md: ✅ / ❌ / N/A
  • Did you write any new necessary tests?: ✅ / ❌ / N/A
  • Did you update Changelog?: ✅ / ❌ / N/A

Additional Information

Summary by CodeRabbit

  • New Features
    • Introduced a new Claude plugin with comprehensive ModelOpt capabilities. The plugin provides essential skills for model quantization, deployment across multiple platforms and engines, comprehensive evaluation frameworks, and intelligent orchestration workflows. Available as version 0.1.0 with complete plugin ecosystem integration.

Signed-off-by: Kai Xu <kaix@nvidia.com>
@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot bot commented Mar 30, 2026

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

@coderabbitai
Copy link
Copy Markdown
Contributor

coderabbitai bot commented Mar 30, 2026

📝 Walkthrough

Walkthrough

Added two new Claude plugin configuration files: a marketplace manifest for modelopt-skills and plugin metadata for modelopt-agent. Both files include version information (0.1.0), descriptive metadata about quantization, deployment, and evaluation capabilities, and contact details.

Changes

Cohort / File(s) Summary
Claude Plugin Configuration
.claude-plugin/marketplace.json, .claude/skills/.claude-plugin/plugin.json
Added plugin marketplace manifest and plugin metadata files defining modelopt-skills marketplace entry and modelopt-agent plugin with version 0.1.0, descriptions covering quantization and deployment, and owner/author contact information.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

🚥 Pre-merge checks | ✅ 4
✅ Passed checks (4 passed)
Check name Status Explanation
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title accurately describes the main change: adding Claude Code plugin marketplace manifest files for ModelOpt agent skills discovery and installation.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Security Anti-Patterns ✅ Passed PR contains only JSON configuration files with no Python code changes, so security anti-patterns checks are not applicable.

✏️ Tip: You can configure your own custom pre-merge checks in the settings.

✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch kaix/agent-marketplace

Comment @coderabbitai help to get the list of available commands and usage tips.

@github-actions
Copy link
Copy Markdown
Contributor

PR Preview Action v1.8.1

QR code for preview link

🚀 View preview at
https://NVIDIA.github.io/Model-Optimizer/pr-preview/pr-1141/

Built to branch gh-pages at 2026-03-30 17:45 UTC.
Preview will be ready when the GitHub Pages deployment is complete.

@kaix-nv kaix-nv marked this pull request as ready for review March 30, 2026 17:46
@kaix-nv kaix-nv requested review from Edwardf0t1 and mxinO March 30, 2026 17:46
Copy link
Copy Markdown
Contributor

@coderabbitai coderabbitai bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.

Inline comments:
In @.claude-plugin/marketplace.json:
- Around line 8-14: The plugin manifest declares "modelopt-agent" with source
"./.claude/skills" but the skills directory only contains plugin metadata and no
executable skill implementations, so commands like /modelopt-agent:* won't be
discovered; add actual skill files (e.g., TypeScript/JavaScript/Python handlers)
into the "./.claude/skills" folder that export the declared commands and
handlers referenced by the plugin manifest (ensure function/command names and
routes match the manifest and plugin.json), include proper entry points and
registration code so the runtime can discover commands, and update the package
manifest if needed to reference the new skill files.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 86e71c35-d8e4-4a64-b5fa-9358f048d2f0

📥 Commits

Reviewing files that changed from the base of the PR and between 48da281 and d7e59cc.

📒 Files selected for processing (2)
  • .claude-plugin/marketplace.json
  • .claude/skills/.claude-plugin/plugin.json

Comment on lines +8 to +14
"plugins": [
{
"name": "modelopt-agent",
"source": "./.claude/skills",
"description": "AI agent skills for model optimization: quantization (PTQ), deployment (vLLM/SGLang/TRT-LLM), evaluation (NEL), and end-to-end orchestration.",
"version": "0.1.0"
}
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟠 Major

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# Verify that skill definition files exist under the declared source path
# (excluding plugin metadata-only files).

set -euo pipefail

echo "Declared source path from marketplace:"
rg -n '"source"\s*:\s*"\./\.claude/skills"' .claude-plugin/marketplace.json

echo
echo "Files under .claude/skills:"
fd . .claude/skills

echo
echo "Potential skill definition files (excluding plugin metadata):"
fd . .claude/skills \
  | rg -n -v '\.claude/skills/\.claude-plugin/plugin\.json$|/\.claude-plugin/|README|LICENSE' || true

Repository: NVIDIA/Model-Optimizer

Length of output: 235


🏁 Script executed:

#!/bin/bash
# Check if .claude/skills directory exists and verify the broader plugin structure

set -euo pipefail

echo "=== Checking .claude/skills directory existence ==="
if [ -d "./.claude/skills" ]; then
  echo "✓ Directory exists"
  echo "Listing contents:"
  ls -la "./.claude/skills" || echo "Cannot list contents"
else
  echo "✗ Directory does NOT exist"
fi

echo
echo "=== Checking for .claude directory structure ==="
if [ -d "./.claude" ]; then
  echo "Contents of .claude directory:"
  find ./.claude -type f | head -20
else
  echo "./.claude directory does not exist"
fi

echo
echo "=== Searching for skill/command definitions elsewhere ==="
# Search for potential skill definition patterns
rg -l '"name".*"modelopt"' --type json --type yaml --type md | head -20 || echo "No matches found"

echo
echo "=== Checking marketplace.json file completely ==="
cat -n .claude-plugin/marketplace.json

Repository: NVIDIA/Model-Optimizer

Length of output: 1195


🏁 Script executed:

#!/bin/bash
# Check the plugin.json metadata to understand the declared skills

echo "=== Contents of .claude/skills/.claude-plugin/plugin.json ==="
cat -n ./.claude/skills/.claude-plugin/plugin.json

echo
echo "=== Check for any .ts, .js, .py files that might define skills ==="
fd -e ts -e js -e py . .claude/skills || echo "No such files found"

echo
echo "=== Check repository root for any skills or commands definitions ==="
rg -l "quantization|deployment|vLLM|SGLang" --type json --max-count 5 || echo "No matches"

Repository: NVIDIA/Model-Optimizer

Length of output: 636


Plugin declares features but contains no skill implementations.

The .claude/skills directory exists but contains only metadata in .claude-plugin/plugin.json. There are no actual skill definition files (.ts, .js, .py, etc.) implementing the promised features (quantization, deployment, evaluation, orchestration). Commands like /modelopt-agent:* will not be discoverable after installation, making the plugin non-functional.

🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.

In @.claude-plugin/marketplace.json around lines 8 - 14, The plugin manifest
declares "modelopt-agent" with source "./.claude/skills" but the skills
directory only contains plugin metadata and no executable skill implementations,
so commands like /modelopt-agent:* won't be discovered; add actual skill files
(e.g., TypeScript/JavaScript/Python handlers) into the "./.claude/skills" folder
that export the declared commands and handlers referenced by the plugin manifest
(ensure function/command names and routes match the manifest and plugin.json),
include proper entry points and registration code so the runtime can discover
commands, and update the package manifest if needed to reference the new skill
files.

@codecov
Copy link
Copy Markdown

codecov bot commented Mar 30, 2026

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.14%. Comparing base (48da281) to head (d7e59cc).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #1141      +/-   ##
==========================================
- Coverage   70.17%   70.14%   -0.03%     
==========================================
  Files         230      230              
  Lines       26053    26053              
==========================================
- Hits        18282    18276       -6     
- Misses       7771     7777       +6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Copy link
Copy Markdown
Contributor

@Edwardf0t1 Edwardf0t1 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it would be better to add this after we’ve finished polishing the skills within our team.

"description": "NVIDIA ModelOpt agent skills for model quantization, deployment, and evaluation.",
"owner": {
"name": "NVIDIA ModelOpt",
"email": "modelopt@nvidia.com"
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

qq: is this email valid?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants