Add AGENTS.md to this repository for coding agent onboarding

## Issue Description

This repository is missing an `AGENTS.md` file at the root. That file is used to **onboard a coding agent (like GitHub Copilot's automation agents)** so it can work efficiently and correctly in this codebase.

Providing high-quality, explicit repository instructions dramatically reduces:
- Broken or failing CI / build attempts
- Wasted exploration (grep/find excessive searches)
- Incorrect environment setup
- PR churn due to missing context

## Required Action

Create a new `AGENTS.md` file in the repository root with the following high‑level sections (in this order):

1. **Overview** – What the project does, main technologies, primary entry points.
2. **Code layout** – Bullet list of important files / folders with brief purpose (Python modules, infra/Bicep, workflows, scripts, config). Omit trivial or generated files.
3. **Running the code** – Step‑by‑step environment setup (preferred Python version / tool, how to install deps, how to run main scripts). Prefer using GitHub Models if applicable over Azure-only model configs.
4. **Running the tests** – Exact commands to run the test suite (include any required env vars, virtualenv activation, or formatting/lint steps if enforced by CI).
5. **Upgrading Python dependencies** – How to add or upgrade a dependency (e.g., edit `requirements.txt` and run a pin/update command). Include any tools (pip, uv, etc.).
6. **Release / build / deployment notes** (if applicable) – Any scripts, workflows, or manual steps the agent must respect.
7. **Conventions & gotchas** – Linting rules, code style, required env vars, secrets handling, common pitfalls.
8. **Validation checklist** – A short ordered list the agent should follow before opening a PR (e.g., tests pass locally, formatting applied, no stray debug prints, etc.).
9. **When to search** – A closing note instructing the agent to trust `AGENTS.md` and only perform source-wide searches if information is missing or incorrect.

## Authoring Guidelines (Summarized from canonical AGENTS.md guidance)

- Be explicit: Prefer concrete commands over prose (e.g., `pytest -q` not “run tests”).
- Verify every command before including it.
- List required environment variables and how they’re obtained.
- Document any non-obvious build or test ordering.
- Surface known flaky areas, hacks, or TODO hotspots if they affect modification strategy.
- Keep it concise but information-dense—optimize for an automated agent, not a marketing audience.

## Example Skeleton

```markdown
# AGENTS.md

## Overview
<1–3 paragraph summary>

## Code layout
- `agent.py` – Entry point for ...
- `checks/` – YAML definitions for automated repo code checks
- `infra/` – Bicep templates for infra provisioning
- `.github/workflows/` – CI/CD workflows
...

## Running the code
1. Create virtual environment: `python -m venv .venv && source .venv/bin/activate`
2. Install deps: `pip install -r requirements.txt`
3. Run main script: `python agent.py --help`

## Running the tests
```bash
source .venv/bin/activate
pytest -q
```

## Upgrading Python dependencies
Edit `requirements.txt`, then: `pip install -r requirements.txt` and run tests.

## Validation checklist
- [ ] Virtualenv active
- [ ] `pytest -q` passes
- [ ] No debug prints
- [ ] Updated docs if behavior changed

## When to search
Only search the codebase if required info is missing or appears outdated.

## Acceptance Criteria

- `AGENTS.md` exists at repository root
- Contains all required sections above
- Commands verified and minimal
- Provides actionable, agent-focused guidance

Adding this file will significantly improve automated contribution quality.

---
*This issue was automatically created by the GitHub Repo Maintainer Agent.*