AI-Native .NET Bootstrap

Turn any .NET repository into an AI-native development environment.

What is this?

A skill that analyzes a .NET repo and scaffolds AI infrastructure. Two tiers:

• Required — AGENTS.md + CI feedback loop
• Optional — Specialized agents, skills, and workflows

Required

1. AGENTS.md

Tells AI agents what your repo is, how to build it, and how to behave.

AGENTS.md    ← Place at repo root (cross-tool standard: Copilot, Cursor, Claude, etc.)

Template: assets/core/AGENTS.md

Contains: repo description, build/test commands, project structure, conventions, agent behavior rules.

2. RL Environment

Lets agents make changes, get CI feedback, and iterate.

Issue assigned → Agent reads AGENTS.md → Makes changes → Pushes PR → CI runs → Reads results → Iterates

Three components:

Component	GitHub Actions	Azure Pipelines	Purpose
Agent build environment	copilot-setup-steps.yml	agent-build-pipeline.yml	Agent can build the repo
pr-build-status skill	pr-build-status.md	(same — supports both)	Agent can read CI failure logs

---

Optional

Agents (multi-step workflows)

Agent	Template	What it does
PR workflow	pr.md + pr/post-gate.md	4-phase PR review + multi-model fix
Test writer	write-tests-agent.md	Routes to correct test project
Learn from PR	learn-from-pr.md	Continuous improvement

Skills (focused capabilities)

Configured (need your repo's commands):

Skill	Template
try-fix	assets/templates/try-fix.md
run-tests	assets/templates/run-tests.md
verify-tests-fail	assets/templates/verify-tests-fail.md

Universal (copy as-is):

Skill	Source
pr-finalize	assets/skills/pr-finalize.md
issue-triage	assets/skills/issue-triage.md
find-reviewable-pr	assets/skills/find-reviewable-pr.md
learn-from-pr	assets/skills/learn-from-pr.md
ai-summary-comment	assets/skills/ai-summary-comment.md

Workflows & Prompts (copy as-is)

File	Source
find-similar-issues	assets/workflows/find-similar-issues.yml
inclusive-heat-sensor	assets/workflows/inclusive-heat-sensor.yml
release-notes	assets/templates/release-notes.prompt.md

Repo Health Monitoring (GitHub only, requires gh-aw)

Component	Source
Setup skill	assets/skills/setup-repo-health-check.md
Health check orchestrator	assets/templates/repo-health-check.md
Investigation worker	assets/templates/repo-health-investigate.md
Dashboard groomer	assets/templates/repo-health-groom.md

Quick Start

1. Add this skill to your Copilot configuration
2. Say: "Bootstrap AI infrastructure for this repo"
3. The skill analyzes your repo and creates everything in .github/

Or for just the required tier: "Set up the core AI infrastructure — just AGENTS.md and CI feedback loop"

How It Works

SKILL.md                    ← Entry point (Required → Optional flow)
├── references/
│   ├── core-setup.md            Required tier guide
│   ├── repo-analysis.md         Detection commands
│   ├── ci-feedback-loop.md      Feedback loop details
│   ├── generating-instructions.md  Instruction templates
│   ├── agents-and-skills.md     Optional tier catalog
│   └── repo-health-check.md     Health monitoring architecture
├── assets/
│   ├── core/              ← Required: AGENTS.md, CI setup
│   ├── templates/         ← Optional: fill in {{PLACEHOLDERS}}
│   ├── skills/            ← Optional: copy unchanged
│   └── workflows/         ← Optional: copy unchanged
└── evals/                 ← Test cases

License

MIT