V0.9.0/dotnet remote testing

[gimlichael]

This pull request updates documentation to clarify and standardize the workflow for running skill evaluation benchmarks, emphasizing the use of the repo-owned PowerShell runner script. The changes improve instructions for both new and existing contributors, ensuring consistency and reliability in skill benchmarking and review.

Standardization of skill benchmarking workflow:

• The preferred entry point for running local skill benchmarks is now the repo-owned scripts/run-skill-benchmark.ps1 PowerShell script, which manages temp workspaces, caches, parallelism, timeouts, and artifact generation, while delegating aggregation and review to Anthropic's installed tools. [1] [2] [3]
• Added explicit instructions and example commands for using the PowerShell runner in README.md and CONTRIBUTING.md, including details on the artifacts produced and the review flow.

Documentation improvements:

• Updated the skill creation/modification workflow to require a full per-skill test using the standardized runner, clarifying what constitutes a full test and the expected review process. [1] [2]
• Added a new example for installing the dotnet-remote-testing skill via npx skills add in the skill installation section.

[greptile-apps]

Greptile Summary

The PR adds the dotnet-remote-testing skill and standardizes measured skill evaluation around the repository-owned PowerShell benchmark runner.

• Adds Docker-based remote .NET test discovery, planning, execution, result parsing, caching, validation, and evaluation fixtures.
• Adds cross-platform benchmark orchestration with bounded parallelism, timeouts, shared caches, grading, aggregation, and review artifacts.
• Updates contributor guidance, skill documentation, installation instructions, and release notes.

Confidence Score: 5/5

The PR appears safe to merge because no blocking failure remains from the previously reported issues.

No blocking failure remains.

Important Files Changed

Filename	Overview
scripts/run-skill-benchmark.ps1	Adds the repository benchmark orchestrator and now creates an executable POSIX dotnet shim alongside the Windows shim, resolving the previously reported Unix instrumentation failure.
skills/dotnet-remote-testing/scripts/remote-test.cs	Implements remote-test discovery and Docker execution; cached release arrays are now reconstructed into the object shape expected by the release parser.
skills/dotnet-remote-testing/FORMS.md	Defines the remote-testing inputs and now exposes the runner-computed project target as a selectable recommended choice.
scripts/test-run-skill-benchmark.ps1	Adds benchmark-runner validation, including checks for the POSIX shim and its executable permission.
skills/dotnet-remote-testing/scripts/test-remote-testing.ps1	Adds the PowerShell harness for validating remote-testing behavior.
skills/dotnet-remote-testing/SKILL.md	Documents deterministic environment selection, Docker execution, failure classification, and cleanup behavior.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Skill eval definitions] --> B[Benchmark runner]
  B --> C[Temporary workspace]
  B --> D[Paired candidate and baseline runs]
  D --> E[Grading]
  E --> F[benchmark.json]
  F --> G[Static review viewer]
  H[Remote testing request] --> I[Environment discovery]
  I --> J[Docker image resolution]
  J --> K[Isolated restore, build, and test]
  K --> L[Structured results and cleanup]

Loading

_{Reviews (3): Last reviewed commit: "updated changelog" | Re-trigger Greptile}