agentic-sw-engineering

Master skill for the reflect->capture->distribute learning loop in any repository: its MAIN TASK is to reflect on a session/task and turn what it taught into durable, reusable knowledge, then distribute it. Reflect on what a session revealed, capture it by authoring/updating skills, agents, docs, and a lessons file, and distribute it both to the team (repo-local `.claude/skills/`, committed) and across projects (global `~/.claude/skills/`) via push learnings up, pull skills down, or reconcile, all through the bundled skill-sync leaf. Delegates each step to the correct leaf skill. Triggers on: "reflect", "review session", "what did we do", "session summary", "learnings", "capture lessons", "update lessons", "improve skills", "update docs", "analyze conversation", "learning loop", "create skill", "update skill", "new skill", "write a skill", "global skill", "promote skill", "sync skills", "reconcile skills", "pull global skill", "skills out of sync", "drift audit", "skill drift".

Audit and fix Markdown files for brand-voice quality in the "Agentic Software Engineering using GitHub Copilot" masterclass (audience: software engineers): em dashes, Mermaid label syntax, paragraph length, and topic-specific slash command tables. Trigger phrases: brand voice check, quality check, check writing quality, fix em dashes.

Master skill for regulatory compliance assessment of a product: routes to the matching leaf and covers both the AI-specific and the personal-data regulations. Leaf `eu-ai-act` handles EU AI Act risk-tier classification (prohibited / high-risk / limited-risk transparency / minimal), the Article 50 transparency-disclosure pattern for consent screens and Terms of Service, and stakeholder/owner checklists. Leaf `dsgvo` handles DSGVO / GDPR personal-data compliance: lawful basis, data-subject rights, privacy notice, processor (Art. 28) agreements, records of processing, DPIA, breach notification, third-country transfers, and the compliance-check checklist. Use when asked to "check compliance", "compliance checklist", "EU AI Act compliance", "is our AI high-risk", "Article 50 transparency", "DSGVO compliance", "GDPR compliance", "check for DSGVO", "are we GDPR compliant", "do we need a privacy notice / Datenschutzerklärung", "lawful basis", "data subject rights", "data processing agreement", "AVV", "DPIA", "data br

Author a single demo or lab guide for a hands-on class of any audience, in one of three styles chosen by audience: prompt-recipe (business/maker, the five-part step model), guided (initial project setup), or walkthrough (code, Microsoft-Learn style with run commands and expected output). Places the file at the module root as demo-NN-slug.md or lab-NN-slug.md, scaffolds any runnable project and companion assets, and updates the folder TOC. Peer skill the create-class master delegates to; also invocable directly. Trigger phrases: create a guide, write a demo, author a lab, demo guide, lab guide, prompt recipe, walkthrough guide, learn style guide, run-verified guide.

Turn a topic readme into presentation slides for a hands-on class, via one of two first-class paths chosen per class. Gamma path: analyse a topic readme and emit Gamma-ready slide spec files (frontmatter plus heading, bullets, optional Mermaid) into the module pptx/ folder, with a merge-slides sub-capability that combines multiple Gamma PPTX exports to beat Gamma's ~10-item-per-call cap. PPTX path: generate a local .pptx deck directly with python-pptx (via the pptx skill) for classes that do not use Gamma. The master create-class skill delegates here; also invocable on its own. Trigger phrases: create slides, topic to slides, gamma slides, pptx deck, slide spec, merge slides, generate deck, slides from readme, build a deck.

Enrich a module README with a use-case intro, a topic-specific slash-command table, and key documentation links, adapted to the module's audience (developer, low-code maker, or business user). Peer skill that the create-class master delegates to by name; also invocable directly. Trigger phrases: enrich module readme, module readme, use case intro, slash command table, key links, teaching content, enhance for business owner, business audience, adapt readme for business users, topic-specific slash commands.

Scaffold, evaluate, render, update, show, test, and publish the class progress dashboard for a course repo. A static web app plus Python evaluators plus a GitHub Pages workflow that report per-module completeness (theory, demos, labs, slides). Optional per class. Peer skill the create-class master delegates to; also invocable directly. Trigger phrases: class dashboard, module status, whats missing, setup dashboard, show status, update class status, test dashboard, publish class dashboard, github pages dashboard, class progress.

Conceive and structure a course's module and topic layout and its master table of contents: name topics so they earn attendance, order modules and topics to build knowledge, keep the master TOC in the author's bare-keyword-link format, keep each title in sync across the TOC, the topic H1, and the module table, size modules so they feel substantial, and keep each capability in exactly one module. Peer subskill of create-class, used before and during scaffolding and whenever restructuring an existing course. Trigger phrases: conceive the toc, design the module structure, order the modules, name a topic, retitle a topic, master toc format, toc bullets, module overlap, thin module, split a module, swap modules, agenda, course outline, engaging titles, restructure demos.

Guide the user through running and verifying any code demo in this repo, then fold what worked back into the demo guide. Auto-detects the stack (.NET, Python, Node), prepares the environment (config, deps, build), and hands the user the exact terminal commands. Does not run interactive or long-running scripts directly. Errors are read from logs and resolved collaboratively. Delegates to run-foundry-demo when the demo is an Azure AI Foundry / Agent Framework demo. Triggers on: "run demo", "run the demo", "run this demo", "start demo", "execute demo", "verify code demo", "evaluate code demo", "run dotnet demo", "run python demo", "run node demo", "does the demo work".

Guide the user through running a Foundry demo in demos/02-agents. Prepares the environment (config, venv, deps) and gives the user the terminal commands to run. Does not run interactive scripts directly. Errors are discussed and resolved collaboratively. Triggers on: "run demo", "run the demo", "run this demo", "start demo", "execute demo".

Run a demo or lab guide as a real user in a live browser or desktop app, observe what actually works, then rewrite the guide from that. Two backends the user picks per run: Claude Code desktop's native in-app browser (simplest, no extra stack) or Midscene plus Qwen3-VL vision automation (cheap, drives Claude Desktop or any native app). The browser and desktop counterpart to run-demo (code) and run-foundry-demo (Foundry). Peer skill the create-class master delegates to; also invocable directly. Trigger phrases: run guide in browser, evaluate guide, test guide as a user, drive claude desktop, in-app browser guide test, midscene guide test.

Bootstrap a new course or masterclass repository and generate its module, topic, and lab layout. Interactive: asks a few questions (class title, audience, module count or outline, whether labs are needed), then creates the base skeleton (.claude/, demos/, labs/, root CLAUDE.md) and numbered module folders. Can bootstrap from an outline file (## Module N: headings become module folders, bullets become numbered topic subfolders) or purely interactively, at flat or deep-nested depth. Standalone skill; also called by the create-class master. Trigger phrases: scaffold a class, scaffold a new class, bootstrap course layout, set up base repo, new class, generate modules from outline, create demos and labs folders, scaffold module, add topic subfolders, bootstrap course repo.

Master skill for authoring a hands-on course or masterclass repository end to end, for any audience (developer, low-code maker, business user). It is the single entry point that delegates to standalone peer skills: module-toc-conception, scaffold, create-teaching, create-guide, create-slides, dashboard, run-demo, run-foundry-demo, run-guide-browser, and the repo-local brand-voice-* skill. Use when building or maintaining any course repo organized as numbered modules under demos/ and labs/. Trigger phrases: create a class, build a class, author a course, scaffold a new class, bootstrap course layout, generate modules from outline, conceive the toc, order the modules, name a topic, engaging titles, restructure demos, write a demo, author a lab, create slides, topic to slides, enrich module readme, class dashboard, module status, whats missing, run demo, verify code demo, run guide in browser, evaluate guide, brand voice check, quality check, business class, maker class, code class.

Master skill for SEO auditing, optimization, and Core Web Vitals tracking on any web project. Delegates to the correct narrow leaf skill. Use when auditing SEO health, updating meta descriptions or title tags, adding structured data (JSON-LD), comparing audit snapshots, setting up performance monitoring, optimizing for AI/GEO citation, writing AEO content blocks, or planning programmatic SEO pages. Triggers on: "SEO audit", "meta description", "structured data", "schema markup", "title tags", "open graph", "core web vitals", "LCP", "CLS", "INP", "seo compare", "seo optimize", "ranking", "sitemap", "GEO", "AI citation", "AI Overviews", "llms.txt", "programmatic SEO", "pSEO", "answer engine", "featured snippets", "E-E-A-T", "content patterns".

Consolidated Angular conventions for components, dependency injection, forms, HTTP, routing, and signals. Use when working on Angular applications and you need modern Angular guidance or routing into a specific Angular concern. Triggers on component creation, inject(), signal forms, httpResource(), guards, route configuration, computed(), linkedSignal(), and Angular architecture.

Deploy container images from ACR to Azure Container Apps — handles create-or-update, external ingress, environment variables, and managed identity registry authentication. Generates GitHub Actions workflow. Keywords: azure container apps, aca, container app, deploy, ingress, environment variables, managed identity, acr pull.

Build Docker images for each service defined in deploy.json and push them to Azure Container Registry. Generates GitHub Actions workflow for CI. Keywords: acr, azure container registry, docker build, docker push, github actions, container image, build pipeline.

Deploy a .NET (or Node/Python) API to Azure App Service using OIDC Workload Identity Federation and zip deploy via GitHub Actions. Keywords: azure app service, app service, dotnet deploy, web app, zip deploy, oidc, workload identity, github actions, api deploy.

Deploy a static UI to Azure Static Web Apps and a backend API to Azure App Service using OIDC Workload Identity Federation. Orchestrates three subskills — wif (identity setup), swa (static UI deploy), appservice (API deploy). Use when deploying any service to Azure via GitHub Actions. Keywords: azure static web apps, azure app service, oidc, workload identity, github actions, deploy.

Deploy a static UI (Angular, React, Vue, Hugo) to Azure Static Web Apps using OIDC Workload Identity Federation. Key insight: azure/static-web-apps-deploy@v1 cannot use the OIDC session directly — the SWA deployment token must be retrieved via Azure CLI after login. Keywords: azure static web apps, swa, oidc, workload identity, angular deploy, static site, github actions.

One-time setup of Azure Workload Identity Federation for GitHub Actions — creates managed identity, assigns AcrPush and Contributor roles, creates federated credential. Run before any deployment. Keywords: workload identity federation, WIF, OIDC, managed identity, federated credential, github actions azure auth, passwordless.

Use this skill to review code. It supports both local changes (staged or working tree) and remote Pull Requests (by ID or URL). It focuses on correctness, maintainability, and adherence to project standards.

Master skill for turning a raw product idea OR existing technical/use-case docs into a polished marketing pitch deck. Optionally shapes a 3-line idea into a structured brief first, then produces a long outline plus a 6 to 7 card Gamma deck, exported to PPTX and per-slide PNG images. Delegates to idea-to-pitch, pitch-sources, pitch-outline, pitch-slides, pitch-brand-voice, pitch-gamma, and pitch-export. Use when the user invokes /create-product-pitch, or says "create a pitch", "idea to pitch", "shape this idea", "product idea pitch", "pitch deck", "marketing pitch", "Gamma pitch", "turn this into slides", "make a pitch out of X", or "create slides for a product".

Ensure .NET/C# code meets best practices for the solution/project. Delegates CLI commands to the dotnet-cli sub-skill.

Stage 1 of the idea-to-pitch flow. Turns a 3-to-many-line product idea into the structured product brief that /create-product-pitch consumes. Runs a ledger-based discussion loop, asks batched clarifying questions, records every decision to an idea-ledger.md, then writes a product brief (readme.md + usecases.md) under the product's own directory. Use when the user invokes /idea-to-pitch, or says "idea to pitch", "turn my idea into a pitch", "shape this idea", "help me develop a product idea", or "I have an idea for a product".

Create Mermaid diagrams for flowcharts, sequences, ERDs, and architectures. Masters syntax for all diagram types and styling. Use PROACTIVELY for visual documentation, system diagrams, or process flows.

Use when tasks involve reading, creating, or reviewing PDF files where rendering and layout matter; prefer visual checks by rendering pages (Poppler) and use Python tools such as `reportlab`, `pdfplumber`, and `pypdf` for generation and extraction.

Use this skill any time a .pptx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates, layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx filename, regardless of what they plan to do with the content afterward. If a .pptx file needs to be opened, created, or touched, use this skill.

Automate creation of Azure DevOps workload identity federation service connections with deployment metadata from deploy.json. Use when users need to create or delete Azure service connections with workload identity federation for secure, passwordless authentication.

Migrate outdated Microsoft Agent Framework Python samples to the GA v1+ API. Use when an existing MAF Python project uses the old package (agent-framework-azure-ai), AzureAIClient, async context managers for client/agent, azure.identity.aio credential, result.text access pattern, or missing AZURE_ env var prefix. Produces an updated requirements.txt, main.py, and .env template aligned to the GA agent-framework package.

Microsoft Agent Framework (MAF) Python reference. Use when creating, configuring, or explaining MAF Python agents, selecting the correct packages, understanding the FoundryChatClient/Agent API, managing credentials with AzureCliCredential, or setting up Azure AI Foundry projects. Covers GA v1+ patterns, package names, imports, env vars, and key rules. Subskills available: maf-update (for migrating outdated MAF samples).

Get logs from the latest Azure DevOps pipeline run using patterns that actually work.

Automate the import and execution of Azure DevOps pipelines from YAML files using deployment metadata. Use this when users need to import pipelines to Azure DevOps, run them, and fix any errors that occur during execution.

Automate browser interactions and tasks using CLI or Python. Use when testing web applications, capturing screenshots, extracting data, automating workflows, testing forms, or running autonomous browser agents. Supports local (Chromium, real Chrome) and cloud (remote) browsers. Keywords: browser automation, web scraping, screenshot, form submission, web testing, browser agent, automation workflow.

Create, edit, analyze, and validate Word documents (.docx files) programmatically. Use when generating reports, creating templates, automating document generation, editing existing documents with tracked changes, or converting documents to other formats. Supports JavaScript (docx-js) for creation and XML editing for fine-grained control. Keywords: create docx, generate word document, edit docx, document automation, word template, tracked changes, docx conversion.

Builds a single PDF (and optional EPUB) from repo Markdown using Pandoc and uploads as a CI artifact. Use when generating documentation PDFs, creating ebook distributions, building PDF from markdown sources, or automating document distribution. Works with GitHub Actions workflow automation and supports custom ordering via .bookorder file.

Create, edit, analyze, and validate PowerPoint presentations (PPTX files). Use when asked to build presentations from scratch, modify existing templates, visualize presentation content, convert slides to images, or verify design quality. Supports text extraction, slide thumbnail generation, design validation, and QA workflows. Requires python-pptx, markitdown, pptxgenjs, LibreOffice, and Poppler.

Help users discover and install specialized skills that extend GitHub Copilot capabilities. Use when users ask how to do something, whether a skill exists for a task, or want to find tools and workflows. Covers skill discovery, installation, and capability extension. Keywords: find skill, search skill, install skill, extend capabilities, discover tools, agent skill.

Create distinctive, production-grade frontend interfaces with exceptional aesthetic direction and full accessibility compliance. Use when building components, pages, or applications with clear visual identity. Covers design thinking, typography, color systems, animation, responsive layouts, and WCAG 2.2 accessibility standards. Supports HTML/CSS, React, Vue, Angular, and modern JavaScript frameworks. Keywords: design frontend, build UI, create component, style page, accessible design, design system.

Deploy OpenClaw AI assistant on Raspberry Pi 4 with Node.js 22+. Use when setting up personal AI assistant, deploying AI chatbot on Pi, configuring voice-enabled AI agent, or automating Raspberry Pi AI setup. Covers hardware requirements, OS installation, SSH setup, and full deployment workflow on 64-bit Raspberry Pi OS Lite.