amondejar-oss 

One class per endpoint for ASP.NET Core Minimal APIs. Binding, validation and OpenAPI stay native; authorization is fail-closed; analyzers catch route mistakes at build time.

Typed HTTP calls to Keel.Endpoints endpoint classes in integration tests: the route, its parameters and the body come from the endpoint itself, not from hand-written URLs.

FluentValidation for Keel.Endpoints and any Minimal API: an ordinary endpoint filter that runs the registered IValidator<T> for each argument and returns the same 400 problem details as .NET 10 validation.

dotnet new templates for Keel.Endpoints: keel-api creates an API (optionally Native AOT), keel-slice adds one operation as a folder with its contract and endpoint.

The SpecTrace CLI (dotnet spectrace): traceability report, spec change review (spectrace.lock), MCP server for AI coding agents, import from Spec Kit and OpenSpec, and xUnit test scaffolding.

The [Implements] and [Verifies] attributes that trace Spec-Driven Development requirements to code and tests. Used by the SpecTrace analyzer.

Spec-Driven Development traceability for .NET. Generates a typed requirements catalog from Markdown specs and reports, at build time, requirements that are not implemented, not tested or changed after review.... More information

Observability for Keel.Endpoints: tags ASP.NET Core's request spans and http.server.request.duration with the endpoint's operation and feature, counts validation errors per field, and optionally traces and... More information

Safe retries for POST and PATCH in Keel.Endpoints and any Minimal API: with an Idempotency-Key header, the first response is stored and retries get it back without running the handler again. Per-user keys, 422... More information

dotnet keel diff: compares the api.manifest.json of two git refs or two files and lists breaking, security and compatible changes, as text, Markdown (for pull request comments and release notes) or JSON. Exits... More information

API versioning for Keel.Endpoints on top of Asp.Versioning: a version per folder (Features/V2/...), unique operation names per version, and [ApiVersion] on endpoint classes. Versioned routes, version headers... More information

Keel.Endpoints endpoint classes as MCP (Model Context Protocol) tools for AI agents, on the official MCP C# SDK: opt-in per endpoint, input and output schemas from the request and response contracts, and every... More information

Safe actions for AI agents on Keel.Endpoints MCP tools: the server asks the user to confirm before a destructive or sensitive tool runs (MCP elicitation, stateless), with a signed confirmation bound to the... More information

AI cost and behavior snapshots for .NET apps that use IChatClient. Record the model's responses once and replay them in CI for free; measure calls, tokens and tool calls per test scenario; compare them with a... More information

dotnet tokenledger: compares the TokenLedger baselines of two git refs and lists, per scenario, the regressions, behavior changes and improvements in model calls, tokens, tools and estimated cost, as text,... More information

TokenLedger for ASP.NET Core integration tests: attributes each model call to the HTTP endpoint that made it (POST api/support/chat makes 3 calls), so baselines and pull request reports show the cost per... More information

Build MCP Apps with Razor components: write the interactive UI that Claude, VS Code and other MCP hosts show in the chat as .razor files, with @onclick and @bind, rendered and run on your server.

Checks the agent skills (SKILL.md) that your NuGet packages bring in on every build: prompt injection, hidden Unicode, remote code execution, secret exfiltration, and skills that are new or changed since they... More information

dotnet new agentic-api: an ASP.NET Core API ready for AI agents. Keel endpoints as MCP tools with user confirmation, a McpRazor review panel inside Claude and Copilot, SpecTrace requirements traced to code and... More information

Audits the agent skills (SKILL.md) that arrive with your NuGet packages and your repository: finds prompt injection, hidden Unicode, remote code execution, secret access and exfiltration, and locks every skill... More information