AL Development Collection for GitHub Copilot

The AL-Development-Collection-for-GitHub-Copilot repository is a community-oriented toolbox designed to move GitHub Copilot from occasional help to a reproducible, auditable part of AL development for Business Central. This is a first public version — created by and for the technical community — and it’s intentionally opened for suggestions and contributions. If you work with AL and want to standardize how AI assists in spec generation, code scaffolding and tests, this repo gives you a practical starting point.

Key capabilities

The collection packages prebuilt components that accelerate common AL tasks: templates for specs and codeunits, unit test skeletons, chat modes focused on specific developer activities and executable workflows that automate repetitive steps. The goal is simple: reduce onboarding friction for AI-driven workflows and make outcomes predictable across teams.

Technical approach: three-layer architecture

The project applies an AI-Native Instructions Architecture organized in three layers:

• Markdown Prompt Engineering: clear, structured instructions that define the task, constraints and examples as a human-readable contract.
• Agent Primitives: reusable chat modes and workflows (e.g., generate spec, scaffold codeunit, create tests) that encapsulate AL best practices as executable primitives.
• Context Engineering: mechanisms like applyTo patterns etc.

Technically, this moves knowledge from individual developers into the repo: prompts become versioned artifacts, and workflows can be reviewed and improved collaboratively.

How to start (recommended workflow)

A pragmatic adoption path:

• Clone the repo into a sandbox workspace and run the example workflows to understand the mechanics.
• Customize the base instructions (naming rules, performance patterns, test conventions) to match your team’s standards.
• Run one end-to-end flow (spec → scaffold → unit test), measure time and quality impact, then iterate.

Keep human validation gates where business logic, pricing or external integrations are involved — automation is an accelerator, not a replacement for review.

Community note — first version and open for contributions

This repository is built by and for the technical community. It is a first public version and intentionally lightweight so it’s easy to try and extend. Your suggestions, bug reports and pull requests are welcome: the project aims to evolve through community feedback and real-world usage. Contribute examples, improve primitives, add new workflows or propose changes to the instruction files — collaborative iteration is how this becomes truly useful at scale.

Risks and recommendations

• Do not adopt blindly: automated workflows multiply both gains and mistakes. Use validation gates for critical paths.
• Localize and enforce conventions: adapt instructions to your organization’s naming, performance and security policies before broad rollout.
• Measure before you generalize: start with a small scope and evaluate ROI (time saved, quality of PRs, reduced rework) before expanding.

References

• AL-Development-Collection-for-GitHub-Copilot (GitHub)
• AI-Native Instructions Architecture — adaptación práctica al desarrollo AL (documentación del proyecto)

Conclusion

This collection is a thoughtful approach to enhance the way teams in AL utilize Copilot: it transforms spontaneous prompts into structured instructions and reusable elements. While it may not solve every challenge, by embracing innovation — personalizing guidelines, ensuring human oversight in critical areas, and continuously refining based on feedback — it truly has the potential to boost productivity significantly.

The repository is by the community and for the community; test it, improve it, and help shape the next versions.

Next steps

Integrate use ofMCP tools from tech community like:

• JeremyVyska/bc-code-intelligence-mcp: BC Code Intelligence MCP Server – Persona-driven workflow orchestration for Business Central development. Provides 16+ MCP tools, layered knowledge system, and intelligent BC pattern analysis through Model Context Protocol.
• SShadowS/al-objid-mcp-server: MCP server for AL Object ID Ninja
• waldo1001/waldo.BCTelemetryBuddy

Recordad esto porque ayuda mucho

Suscríbete al canal (anima y da ese empujón a todo esto).

Pulsa «like» si te ha gustado.

Si no quieres perderte nada, ya sabes, pulsa la campana.

En los comentarios déjame cualquier idea, duda, corrección o aportación. Todo será bien bienvenido.

• LinkedIn
• Twitter
• Instagram
• YouTube
• X

Nota:
ES-El contenido de este artículo ha sido generado en parte con la ayuda de IA para revisión, orden o resumen.

El contenido, las ideas ,comentarios ,opiniones son totalmente humanas. Los posts pueden basarse o surge la idea de escribirse de otro contenido se referenciará ya sea oficial o de terceros.

Por supuesto ambas humana e IA pueden contener errores.

Te animo a que en los comentarios lo indiques, para más información accede a la página sobre responsabilidad AI del blog TechSphereDynamics.

EN-The content of this article has been generated in part with the help of IA for review order or summary.

The content, ideas, comments, opinions are entirely human. The posts can be based or arises the idea of writing another content will be referenced either official or third party.

Of course both human and IA can contain errors.

I encourage you to indicate in the comments, for more information go to the page on responsibility AI of the blog TechSphereDynamics.

Original Post https://techspheredynamics.com/2025/10/31/al-development-collection-for-github-copilot-a-practical-toolkit-to-bring-ai-into-your-al-development/