Forecast next month's demand per product from sales history with a statistical baseline, optionally refined by OpenAI/Anthropic/Gemini, and get reorder quantities/dates plus overstock/dead-stock flags. Includes a color-coded admin grid, on-demand/cron scheduling, and tunable settings for lead time, safety stock, and history window.
Adds an AI Content button to PageBuilder and inline AI actions for fields to generate full layouts, copy, and meta fields using OpenAI or Claude. Includes presets, saved prompts, bulk generation, a request audit log, and PageBuilder‑native output; works with Hyvä and Luma.
A priority-ordered roadmap for using AI at speed without chaos: automated deployments, ephemeral environments, E2E and unit tests, static analysis, and security tooling. Explains what “good” looks like, what breaks without each layer, and where to start.
Reusable “skills” and workflow “agents” to extend AI coding assistants (Claude, Copilot, Cursor, etc.), organized under skills/ and agents/, with seamless Dockergento integration. Includes an RFP Analysis Pipeline to extract requirements, generate INVEST user stories with BDD criteria, decompose work items, and produce effort estimates.
Provides a /mcp JSON-RPC endpoint so AI agents can read and write store data, with OAuth 2.1 or bearer tokens, per-tool ACLs, audit logging, rate limiting, and origin allowlists. Includes quick-start setup, CLI commands, and optional sub-modules for orders, catalog, customers, CMS, marketing, reports, and GA4.
Connects Claude, Cursor, Copilot, Gemini, and others to your codebase via an n98-magerun2–backed MCP server, plus curated guidelines and skills for common workflows. Includes an interactive installer, update command, and an extension system that merges project-specific rules and skills from Composer packages.
A collection of 30 agent skills that let AI assistants in Cursor deliver expert guidance across feature implementation, frontend/back‑end work, debugging, performance, and infrastructure. Each skill includes when‑to‑use criteria, step‑by‑step processes, best practices, and links to official docs; MIT‑licensed with contribution guidelines.
Reframes platform selection around data ownership and a self-hosted AI stack, demoing Hyvä + Typesense with agentic content, AI enrichment, and natural-language reporting. Includes reproducible benchmarks on identical hardware-5.2x faster median responses vs Luma-and links to Terraform/Helm for replication.
Claude Code plugin that injects only the rules relevant to your task via a Neo4j-backed RAG pipeline, then blocks risky writes until a plan and tests are approved. Includes quick-start scripts, a FastAPI service, and a CLI, with sub-millisecond retrieval at 10k-rule scale.
AI is rapidly reshaping the way we build, maintain, and scale software. Over the past few weeks, we’ve seen increasing discussions across the community how this impacts daily work, project delivery, and the broader ecosystem.