agentleFS
michaelalbermichaelalber/ai-toolkit/skills/nuget-package-scaffold/SKILL.md

nuget-package-scaffold

Skill3 starsChanged 4 months agoMITView on GitHub

NuGet package creation with CI/CD pipeline setup and test harness. Use when creating new NuGet packages, configuring package metadata, or setting up publish workflows. Do NOT use when the library is internal-only and not intended for NuGet publication; Do NOT use when the target is an application project.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create the solution structure:

    line 36
  2. Distinctive:

    Verify all unit tests pass on every target framework, public API surface is tested, edge cases and error paths are covered.

    line 83

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

NuGet Package Scaffold

"The best libraries are those that disappear into the code that uses them." -- Krzysztof Cwalina, Framework Design Guidelines

Core Philosophy

This skill scaffolds production-ready NuGet packages with proper metadata, testing, CI/CD pipelines, and publish workflows. Every package must meet a quality bar before it reaches any feed.

Non-negotiable constraints:

  1. No publish without tests -- a package without a passing test suite is not a package, it is a liability
  2. Semantic versioning is law -- every version bump must communicate intent: breaking, feature, or fix
  3. Metadata is not optional -- description, license, authors, repository URL, and README are required before first publish
  4. Multi-target when possible -- support the broadest reasonable set of target frameworks for your consumers
  5. Deterministic builds -- the same source must produce the same binary, every time

The full domain principles, knowledge-base lookup protocol, AI discipline rules, anti-patterns, and error-recovery procedures live in references/conventions.md.

Workflow

Six-step pipeline: Scaffold → Configure → Test → Pack → Publish (with iterate back if needed).

Step 1: Scaffold

Create the solution structure:

PackageName/
├── src/
│   └── PackageName/
│       ├── PackageName.csproj
│       └── Class1.cs
├── tests/
│   └── PackageName.Tests/
│       ├── PackageName.Tests.csproj
│       └── Class1Tests.cs
├── .github/
│   └── workflows/
│       ├── ci.yml
│       └── publish.yml
├── PackageName.sln
├── Directory.Build.props
├── README.md
├── LICENSE
└── .editorconfig

Step 2: Configure .csproj Metadata

Required metadata properties checklist:

  • PackageId -- unique identifier on the feed
  • Version -- SemVer-compliant version string
  • Authors -- comma-separated author names
  • Description -- concise summary of what the package does
  • PackageLicenseExpression -- SPDX license identifier (e.g., MIT, Apache-2.0)
  • PackageProjectUrl -- link to the project repository
  • RepositoryUrl -- source repository URL
  • RepositoryType -- typically git
  • PackageReadmeFile -- path to the README to embed in the package
  • PackageTags -- space-separated discovery tags
  • TargetFrameworks -- semicolon-separated TFMs

Multi-target decision: TargetFrameworks: netstandard2.0;net10.0 when pre-.NET 10 consumers need support; TargetFrameworks: net10.0 otherwise. Full property reference: references/csproj-metadata.md.

Step 3: Test

dotnet test --configuration Release

Verify all unit tests pass on every target framework, public API surface is tested, edge cases and error paths are covered.

Step 4: Pack and Verify

dotnet pack --configuration Release --output ./artifacts
dotnet nuget verify ./artifacts/PackageName.1.0.0.nupkg
unzip -l ./artifacts/PackageName.1.0.0.nupkg

Step 5: Publish

# To nuget.org
dotnet nuget push ./artifacts/PackageName.1.0.0.nupkg \
  --api-key $NUGET_API_KEY \
  --source https://api.nuget.org/v3/index.json

# To a private feed
dotnet nuget push ./artifacts/PackageName.1.0.0.nupkg \
  --api-key $FEED_API_KEY \
  --source https://pkgs.dev.azure.com/org/project/_packaging/feed/nuget/v3/index.json

State Block

<nuget-scaffold-state>
step: [scaffold | configure | test | pack | publish]
package_name: [name]
target_frameworks: [frameworks]
version: [version]
publish_target: [nuget.org | private feed | local]
last_action: [what was done]
next_action: [what's next]
blockers: [issues]
</nuget-scaffold-state>

Output Template

TemplateRequired Content
Project ScaffoldCreated files table (path → purpose), initial state block
.csproj ConfigurationConfigured properties list (PackageId, Version, Authors, etc.), Build settings (Deterministic, Source Link, Nullable), state block
CI/CD Pipeline SetupPipeline stages table (Stage / Trigger / Actions: Build=Push/PR, Pack=Tag push, Publish=Tag+approval), state block

Full templates: references/cicd-templates.md.

Discipline, Anti-Patterns & Recovery

AI discipline rules (never publish without tests, validate metadata, multi-target verification, API-surface review), the anti-pattern catalog, and step-by-step error recovery all live in references/conventions.md.

Integration with Other Skills

  • dotnet-vertical-slice -- use to structure the internal architecture, then use this skill to wrap it as a distributable NuGet package
  • TDD skills (tdd-loop, tdd-agent) -- develop the package internals using TDD workflows before progressing to the Pack step; the test project created by this scaffold integrates directly with TDD phase management
  • mcp-server-scaffold (in the edge-ai-robotics-automation-toolkit supplement) -- when building an MCP server that ships as a dotnet tool NuGet package, use this skill for the packaging and mcp-server-scaffold for the server implementation

Reference Files

  • .csproj Metadata Reference -- Complete property reference for package metadata
  • CI/CD Pipeline Templates -- GitHub Actions and Azure DevOps pipeline templates
  • Conventions, Discipline & Recovery -- Domain principles, KB lookups, discipline rules, anti-patterns, error recovery

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.