MMCPMarkets
JA

j0hnnymiller/ai-assisted-software-development

Deployment & DevOps
8 stars Quality 58 Trend 36NOASSERTION
Repository Install

This repo was created for the Leveraging AI for Legacy Code Refactoring session at the 2025 Next Gen AI conference.

Overview

Files suppoting the CODE Training AI Assisted Software Development not included in the public repos.-Course This repo was created for the AIASD CODE Training Course - Config distribution reports differences and blockers with reasons by default; add -Detailed for every selected, in-scope config file or -Repo 'AIASD/Fleetly' to plan one repository. Report boundaries and write safeguards distinguish visibility from permission to copy. - Codebase Explorer combines unfamiliar-code orientation, bounded behavior/dependency tracing, and evidence-based archaeology. It requests only read/search tools, returns findings in chat, and reports unavailable history instead of adding command execution. Use create-codebase-explorer-agent only to create or adapt a profile, not to perform an investigation.

README

ai_generated: true model: “unknown/unknown@unknown” operator: “johnmillerATcodemag-com” chat_id: “0ee42cac-b369-4e34-9193-397eba0fac7a” prompt: | files in the .github/prompts folder should be campared with the canonical skills. If a skill does not exist for the prompt, create a new skill. Prompt files are being deprecated. update the instructions, skills, and agents, responsible for syncing .github files. started: “2026-10-17T18:21:06.8108325-07:00” ended: “2026-10-17T18:37:09.0441419-07:00” task_durations:

  • task: “Prompt-to-skill migration and sync guidance” duration: “00:16:02” total_duration: “00:16:02” ai_log: “ai-logs/2026/10/17/0ee42cac-b369-4e34-9193-397eba0fac7a/conversation.md” source: “user request” revisions:
  • chat_id: “c16f927f-91c4-4a30-aada-8f74b099e547” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Added focused Vision and Scope authoring/review guidance, distinct from feature-level PRD authoring.” ai_log: “ai-logs/2026/10/18/c16f927f-91c4-4a30-aada-8f74b099e547/conversation.md”
  • chat_id: “c16f927f-91c4-4a30-aada-8f74b099e547” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Added Object Role Modeling authoring and validation guidance, distinct from model-to-use-case derivation.” ai_log: “ai-logs/2026/10/18/c16f927f-91c4-4a30-aada-8f74b099e547/conversation.md”
  • chat_id: “c16f927f-91c4-4a30-aada-8f74b099e547” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Documented the unified read-only codebase explorer and clarified its authoring skill.” ai_log: “ai-logs/2026/10/18/c16f927f-91c4-4a30-aada-8f74b099e547/conversation.md”
  • chat_id: “c16f927f-91c4-4a30-aada-8f74b099e547” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Documented scoped skill and Mermaid authoring guidance and corrected slide workflow routing.” ai_log: “ai-logs/2026/10/18/c16f927f-91c4-4a30-aada-8f74b099e547/conversation.md”
  • chat_id: “c16f927f-91c4-4a30-aada-8f74b099e547” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Added focused Agent Host session operating guidance and distinguished harness-specific behavior.” ai_log: “ai-logs/2026/10/18/c16f927f-91c4-4a30-aada-8f74b099e547/conversation.md”
  • chat_id: “0ee42cac-b369-4e34-9193-397eba0fac7a” model: “unknown/unknown@unknown” date: “2026-10-17” change: “Added approved technical-debt and vertical-slice skills, updated four existing skills, and documented the new workflows.” ai_log: “ai-logs/2026/10/17/0ee42cac-b369-4e34-9193-397eba0fac7a/conversation.md”
  • chat_id: “skill-naming-convention-20261018” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Applied approved skill naming decisions and updated active references; preserved workflow behavior and historical provenance.” ai_log: “ai-logs/2026/10/18/skill-naming-convention-20261018/conversation.md”
  • chat_id: “create-implementation-skills-20261018” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Documented the implementation-skill authoring workflow in the skills overview.” ai_log: “ai-logs/2026/10/18/create-implementation-skills-20261018/conversation.md”
  • chat_id: “0ee42cac-b369-4e34-9193-397eba0fac7a” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Separated standalone Marp merging from the course-manifest pipeline and added roadmap-planning and publication-transfer skills.” ai_log: “ai-logs/2026/10/17/0ee42cac-b369-4e34-9193-397eba0fac7a/conversation.md”
  • chat_id: “c16f927f-91c4-4a30-aada-8f74b099e547” model: “unknown/unknown@unknown” date: “2026-10-18” change: “Documented the canonical Marp review workflow, VTI training-index support, and approved Azure hosting skill import from Messer.” ai_log: “ai-logs/2026/10/18/c16f927f-91c4-4a30-aada-8f74b099e547/conversation.md”

AI-Assisted-Software-Development-Course

Files suppoting the CODE Training AI Assisted Software Development not included in the public repos.-Course This repo was created for the AIASD CODE Training Course

Guidance & Instructions

  • Config distribution reports differences and blockers with reasons by default; add -Detailed for every selected, in-scope config file or -Repo 'AIASD/Fleetly' to plan one repository. Report boundaries and write safeguards distinguish visibility from permission to copy.
  • Codebase Explorer combines unfamiliar-code orientation, bounded behavior/dependency tracing, and evidence-based archaeology. It requests only read/search tools, returns findings in chat, and reports unavailable history instead of adding command execution. Use create-codebase-explorer-agent only to create or adapt a profile, not to perform an investigation.
  • Agent Host session guidance covers workspace and harness selection, isolation, permissions, parallel work, validation, and handoff; it distinguishes current harness-specific behavior rather than assuming one runtime configuration applies everywhere.
  • Agent Skill authoring defines entry-point naming, discovery descriptions, resource safety, and validation for repository skills.
  • Mermaid authoring covers semantic colors, non-color cues, and rendering checks. It automatically targets standalone .mmd sources; attach it or load it by task relevance for diagrams embedded in Markdown.
  • Object Role Modeling guides conceptual fact-model creation and validation; use-cases-from-conceptual-model separately derives use cases from a model that already exists.
  • Vision and Scope guidance supports evidence-based system assessment, options, and delivery direction; use PRD authoring for focused product/feature requirements.
  • The Slide Master agent and Course-specific merging, review, pipeline scripts, and instructions are maintained in the Course repository, not this hub.

Useful Keybinding

This keybinding opens markdown files in a side-by-side preview and copies the editor to a new window for easier review:

{
  "key": "ctrl+shift+alt+x",
  "command": "extension.multiCommand.execute",
  "args": {
    "sequence": [
      "markdown.showPreviewToSide",
      "workbench.action.copyEditorToNewWindow"
    ]
  },
  "when": "editorLangId == mdc"
}

This requires the multiCommand extension and is configured to trigger when editing Marp markdown files (.mdc).

Agent Skills

Reusable task workflows are maintained as workspace skills in .github/skills/. Each skill has a SKILL.md entry point and can include supporting scripts or references. Legacy prompts are migration-only; see the migration log.

Skill names use lowercase kebab-case, and each folder name matches the name in its SKILL.md. Related domain and platform skills retain shared prefixes; task workflows use descriptive action names. The approved naming decisions and rename record cover all 80 skills: 50 renamed and 30 kept. A deprecate- prefix is a naming label only; these renames do not disable skills or change their workflows.

Use create-implementation-skills to author or update focused, repository-grounded implementation workflows.

  • Use summarize-transcript-content for VTT transcripts, VTI training indices with nested duration markers, or structured response aggregation. VTI summaries preserve hierarchy, distinguish declared and calculated durations, and flag incomplete or conflicting timing.
  • Use inventory-technology-stack to list the languages, runtimes, frameworks, libraries, container images, and CI actions used in a folder tree, with each technology’s version and reference count (manifest declarations + matching import statements + source files). Its read-only scanner needs only Python 3.11+ and outputs Markdown, JSON, or CSV.
  • Use app-hosting-decision-tree for a citizen-developer Azure app-host selection interview with functional-fit and cost screening. Imported unchanged from Messer, including its East US 2 pricing default. Use databricks-hosting-decisions for a broader evidence-based hosting comparison without a fixed pricing-region default.

AI-Assisted Artifacts & Provenance

This repository enforces provenance and logging for any AI-assisted outputs (code, docs, diagrams, tests, data).

Guidance & Instructions

Custom GitHub Copilot Chat Modes

Custom chat modes are specialized AI assistants that extend GitHub Copilot’s capabilities for specific domains and workflows. Activate with @ in GitHub Copilot chat.

Role-Based Chat Modes (Generated from Personas)

Utility Chat Modes

Custom GitHub Copilot Agents

Instruction-Authoring Skills

For legacy prompt migration history, see the migration log.

Notable Artifacts

  • Claude Code Assisted Software Development: 3-Day Course Outline (outlines/claude-code-assisted-software-development-3-day.md)

    • Detailed instructor-led course outline covering collaboration, implementation, validation, team workflows, and a PR-ready capstone
    • Provenance: Chat log | Summary

  • Claude Code Assisted Software Development: 5-Day Course Outline (outlines/claude-code-assisted-software-development-5-day.md)

    • Detailed five-day course outline that labels every module as the same as, expanded from, or new relative to the three-day course
    • Provenance: Chat log | Summary

  • Claude Code Assisted Software Development: 3-Day Course Outline (4 Hours Per Day) (outlines/claude-code-assisted-software-development-3-day-4-hour.md)

    • Condensed three-day offering with four instructional hours per day and preserved core learning outcomes
    • Provenance: Chat log | Summary

  • Claude Code Assisted Software Development: 5-Day Course Outline (4 Hours Per Day) (outlines/claude-code-assisted-software-development-5-day-4-hour.md)

    • Condensed five-day offering with four instructional hours per day and explicit same, expanded, moved, new, and removed-topic notation
    • Provenance: Chat log | Summary

  • AI-Assisted MFC to WPF Conversions Slides (slides/marp/ai-assisted-mfc-to-wpf-conversions.deck.md)

    • Marp slide deck for planning and executing MFC-to-WPF migration using AI-assisted workflows
    • Covers both whole-application conversion and module-by-module migration patterns
    • Includes architecture mapping, quality gates, estimation heuristics, and presenter notes on every slide
    • Provenance: Chat log | Summary

  • Thursday GE Merged Course Deck (slides/merged/ge/aiasd-311-thursday.ge-draft.md)

    • Manifest-driven combined Marp deck assembled from slides/manifests/ge/aiasd-311-thursday.ge.manifest.md
    • Includes section divider slides, module progression context, and speaker notes coverage across all slide blocks
    • Provenance: Chat log | Summary

  • VTT Content Summarizer Skill (.github/skills/summarize-transcript-content/SKILL.md)

    • Analyzes VTT (Video Text Track) files from class recordings and generates structured summaries
    • Extracts timing information, section breakdowns, key topics, and action items
    • Includes sample files: VTT input | Generated summary
    • Documentation: Usage Guide | Live Demo
    • Provenance: Chat log

  • Technical Debt Remediation Skill (.github/skills/remediate-technical-debt/SKILL.md)

    • Plans and implements bounded debt fixes after approval, with focused tests and verification.

  • Vertical Slice Implementation Skill (.github/skills/implement-vertical-slice/SKILL.md)

    • Implements one approved slice at a time with explicit acceptance coverage and handoff evidence.

  • Implementation Roadmap Planning Skill (.github/skills/plan-implementation-roadmap/SKILL.md)

    • Converts authoritative requirements and dependency maps into phased execution plans with testable exit gates; does not implement work.

  • Blog Publication Transfer Skill (.github/skills/transfer-blog-publication/SKILL.md)

    • Creates reversible publication-format copies using confirmed transformation rules while preserving source posts.

  • IEC 62304 Work Instruction Authoring Skill (.github/skills/author-iec62304-work-instruction/SKILL.md)

    • Drafts an AI-assisted medical-device software work instruction from an approved QMS template and verified sources; it does not make compliance or approval claims.
    • Provenance: Chat log

  • Course Slide Deck Skill (.github/skills/build-course-slide-deck/SKILL.md)

    • Describes course-manifest merged Markdown and PPTX outputs; execution requires the Course repository’s scripts, which are not included in this hub.
    • Documents the sections/decks schema, derived outputs, note validation, and local finalization.
    • Provenance: Chat log

  • Updated Existing Skills

  • VS Code Copilot Agents Overview Slides (slides/marp/vscode-copilot-agents-overview.deck.md)

    • Comprehensive Marp slide deck covering VS Code Copilot Agents ecosystem with interactive workflows
    • 12 slides with detailed speaker notes covering local, background, cloud, and third-party agents
    • Includes decision matrices, hand-off workflows, and practical implementation guidance
    • Provenance: Chat log | Summary

  • GitHub Worktrees Guide Slides (slides/marp/github-worktrees-guide.deck.md)

    • Comprehensive Marp slide deck teaching parallel development with Git worktrees
    • 12 slides covering introduction, essential commands, practical workflows, hands-on exercises, and best practices
    • Includes detailed speaker notes, troubleshooting guidance, and resources for continued learning
    • Provenance: Chat log | Summary

  • Creating Custom Agents Slides (slides/marp/creating-custom-agents.deck.md)

    • Comprehensive Marp slide deck teaching how to create specialized GitHub Copilot custom agents
    • 12 slides covering overview, creation workflows across platforms (GitHub, VS Code, JetBrains, Eclipse, Xcode), configuration, examples, usage, and best practices
    • Includes extensive speaker notes with timing guidance, delivery instructions, audience interaction points, and Q&A preparation
    • Provenance: Chat log | Summary

  • Code Explanation and Analysis Slides (slides/marp/code-explanation-and-analysis.deck.md)

    • Marp slide deck for Section 10 of the AI-Assisted Software Development course covering code explanation and test coverage gap analysis
    • 11 slides covering inline chat (Ctrl+I), right-click explain, test code understanding, coverage report generation (calculator service example), gap identification, prioritized implementation plans, and hands-on exercises
    • Includes comprehensive speaker notes with timing guidance, live demo instructions, audience interaction points, and exercise facilitation tips
    • Provenance: Chat log

  • AI-Assisted Output Slides (slides/marp/ai-assisted-output.deck.md)

    • 10-slide Marp deck on provenance metadata, placement policy, logging workflow, quality gates, and CI enforcement
    • Sourced from .github/instructions/ai-assisted-output.instructions.md
    • Provenance: Chat log

  • CQRS Architecture Slides (slides/marp/cqrs-architecture.deck.md)

    • 12-slide Marp deck on when to use CQRS, core principles, architecture components, consistency strategies, anti-patterns, and migration
    • Sourced from .github/instructions/cqrs-architecture.instructions.md
    • Provenance: Chat log

  • AI-Assisted Software Architecture Slides (slides/marp/ai-assisted-software-architecture.deck.md)

    • 13-slide Marp deck on AI’s role across the architecture lifecycle: ADRs, tradeoff analysis, vertical slices, CQRS/event-driven patterns, diagram generation, governance, risks, and a practical requirements-to-architecture workflow
    • Includes a best practices checklist and comprehensive speaker notes on every slide
    • Provenance: Chat log | Summary

  • Dependency Management Policy Slides (slides/marp/dependency-management-policy.deck.md)

    • 12-slide Marp deck on risk classification, selection criteria, approval workflow, vulnerability SLAs, license compliance, and supply chain security
    • Sourced from .github/instructions/dependency-management-policy.instructions.md
    • Provenance: Chat log

  • GitHub CLI Slides (slides/marp/github-cli.deck.md)

    • 10-slide Marp deck on issue management, PR workflows, Actions monitoring, code review, and CI/CD integration
    • Sourced from .github/instructions/github-cli.instructions.md
    • Provenance: Chat log

  • Business Rules to Vertical Slices (slides/marp/business-rules-to-slices.deck.md)

    • 10-slide Marp deck on analysis workflow, rule types, use case identification, feature boundary tests, and vertical slice design
    • Sourced from .github/instructions/business-rules-to-slices.instructions.md
    • Provenance: Chat log

  • AI-Assisted Dev Overview (slides/marp/ai-assisted-dev-overview.deck.md)

    • 62-slide combined Marp deck merging all 6 individual module presentations with section divider slides
    • Modules: AI Output · CQRS · Dependency Management · GitHub CLI · Business Rules to Slices · Custom Agents
    • Usable as a full-day course or per-module standalone; every slide has comprehensive speaker notes
    • Provenance: Chat log | Summary

  • AIASD Class 311 Monday (slides/marp/aiasd-311-monday.deck.md)

    • Combined Marp deck for the Monday session of Class 311; ~30 slides across 5 modules with section dividers
    • Modules: AI Output · Vertical Slice Architecture · Creating Prompt Files · Dependency Management · Custom Chat Modes
    • Every slide has comprehensive speaker notes; section dividers allow standalone per-module delivery
    • Manifest: slides/aiasd-311-monday.yaml
    • Provenance: Chat log | Summary

  • AIASD Class 311 — Tuesday Session (slides/marp/aiasd-311-tuesday.deck.md)

    • Combined Marp deck for Class 311 Tuesday, generated from manifest slides/aiasd-311-tuesday.yaml
    • Modules: AI Output Standards · CQRS Architecture · GitHub CLI · Business Rules to Vertical Slices · Custom Agents
    • ~35 slides with section dividers and comprehensive speaker notes on every slide
    • Provenance: Chat log | Summary

  • Code Explanation and Analysis Slides (slides/marp/code-explanation-and-analysis.deck.md)

    • Marp slide deck for Section 10 of the AI-Assisted Software Development course covering code explanation and test coverage gap analysis
    • 11 slides covering inline chat (Ctrl+I), right-click explain, test code understanding, coverage report generation (calculator service example), gap identification, prioritized implementation plans, and hands-on exercises
    • Includes comprehensive speaker notes with timing guidance, live demo instructions, audience interaction points, and exercise facilitation tips
    • Provenance: Chat log

  • Legacy Prompt Migration Instructions (.github/instructions/prompt-file.instructions.md)

    • Historical prompt-file format and migration guidance; new reusable workflows must be skills
    • Provenance: Chat log

  • Instruction Validation Report (validation-report-20251015-212137.md)

  • CQRS Architecture Slides (slides/marp/cqrs-architecture.deck.md)

    • Comprehensive presentation explaining Command Query Responsibility Segregation (CQRS) pattern
    • Covers when to use CQRS, core principles, implementation examples, and migration strategies
    • Includes detailed speaker notes for effective delivery
    • Provenance: Chat log

  • Getting Started Checklist (slides/marp/getting-started-checklist.deck.md)

    • Converted 3-slide Marp deck covering phased AI adoption steps and a high-level AI-assisted workflow from requirements to implementation
    • Preserves the extracted workflow visual and adds speaker notes for presentation use in the course slide pipeline
    • Provenance: Chat log | Summary

  • GitHub Copilot Memory Feature (slides/marp/copilot-memory-feature.deck.md)

    • Marp deck explaining the Copilot memory feature, including user, session, and repository memory scopes
    • Covers memory workflow, memory hygiene, and a slide-authoring example for practical application
    • Provenance: Chat log | Summary

  • Vertical Slicing Architecture Introduction (slides/marp/vertical-slicing-architecture-introduction.deck.md)

    • 7-slide Marp deck introducing feature-based architecture, layered versus slice-oriented structure, and the maintainability benefits of localized change
    • Covers developer experience improvements, team collaboration, testing strategy, and the introductory CQRS relationship
    • Includes Mermaid visuals and detailed speaker notes for classroom delivery
    • Provenance: Chat log | Summary

  • Instruction File ApplyTo Patterns Slides (slides/marp/instruction-file-applyto-patterns.deck.md)

    • 13-slide Marp presentation explaining glob pattern matching for instruction file applyTo fields
    • Covers universal patterns, file extension matching, directory-specific patterns, and best practices
    • Includes comprehensive speaker notes (1-3 minutes per slide) with timing, examples, and audience interaction cues
    • Features real-world examples from this repository, common pitfalls, testing strategies, and decision frameworks
    • Provenance: Chat log | Summary

Utility Scripts

This repository includes utility scripts for repository management and maintenance:

AI Instruction Synchronization

Use config distribution for manifest-controlled planning and approved synchronization. The standalone instruction-sync and copy helpers have been retired.

Document Conversion

Pandoc Configuration Files

Standardized document conversion configurations for slides, documentation, and presentations:

  • slides-to-pptx.yaml - Optimized for converting Marp slides to PowerPoint presentations
  • to-pdf.yaml - Professional PDF output for slides and documentation
  • pandoc-defaults.yaml - Comprehensive multi-format configuration
  • PANDOC.md - Complete usage guide with examples and customization options
  • templates/ - PowerPoint reference templates for professional formatting

Quick Examples:

# Convert slide deck to PowerPoint with template support
pandoc --defaults=slides-to-pptx slides/marp/github-worktrees-guide.deck.md -o github-worktrees.pptx

# Convert documentation to PDF
pandoc --defaults=to-pdf README.md -o project-overview.pdf

Note: For professional PowerPoint formatting, create a reference template in the templates/ directory. See templates/README.md for instructions.

View the full README on GitHub

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add j0hnnymiller/ai-assisted-software-development
New to MCP? Start with the mcpskill.org guide →