Prompt Engineering Series — Roadmap

Planned articles and topic map for the Prompt Engineering for GitHub Copilot series

Author

Dario Airoldi

Published

February 20, 2026

This document tracks the planned and published articles in the Prompt Engineering for GitHub Copilot series, organized by Diátaxis category.

🏗️ Folder structure

The series follows the Diátaxis framework with six content folders:

Folder	Diátaxis type	Purpose
`01-overview/`	Orientation	Series entry point and high-level map
`02-getting-started/`	Tutorial	First steps—naming, organizing, and Copilot Spaces
`03-concepts/`	Explanation	Mental models behind each customization mechanism
`04-howto/`	How-to	Task-oriented guides for building and optimizing prompts
`05-analysis/`	Analysis	Case studies and applied multi-agent patterns
`06-reference/`	Reference	Settings, IDE support, and compatibility tables

Number	Title	Status
01.00	The GitHub Copilot customization stack	✅ Published
01.01	Appendix: Copilot Spaces	✅ Published

Number	Title	Status
02.00	How to name and organize prompt files	✅ Published

Number	Title	Status
01.02	How Copilot assembles and processes prompts	✅ Published
01.03	Understanding prompt files, instructions, and context layers	✅ Published
01.04	Understanding agents, invocation, handoffs, and subagents	✅ Published
01.05	Understanding skills, hooks, and lifecycle automation	✅ Published
01.06	Understanding MCP and the tool ecosystem	✅ Published
01.07	Understanding LLM models and model selection	✅ Published
01.08	Chat modes, Agent HQ, and execution contexts	✅ Published

Number	Title	Status
03.00	How to structure content for prompt files	✅ Published
04.00	How to structure content for agent files	✅ Published
05.00	How to structure content for instruction files	✅ Published
06.00	How to structure content for skill files	✅ Published
07.00	How to create MCP servers for Copilot	✅ Published
08.00	How to optimize prompts for specific models	✅ Published
08.01	Appendix: OpenAI prompting guide	✅ Published
08.02	Appendix: Anthropic prompting guide	✅ Published
08.03	Appendix: Google prompting guide	✅ Published
09.00	How to use agent hooks for lifecycle automation	✅ Published
09.50	How to leverage tools in prompt orchestrations	✅ Published
10.00	How to design orchestrator prompts	✅ Published
11.00	How to design subagent orchestrations	✅ Published
12.00	How to manage information flow during prompt orchestrations	✅ Published
13.00	How to optimize token consumption during prompt orchestrations	✅ Published
14.00	How to use prompts with the GitHub Copilot SDK	✅ Published

Number	Title	Status
20	How to create a prompt orchestrating multiple agents	✅ Published
21.1	Prompt creation multi-agent flow — Implementation plan	✅ Published
22	Prompts and markdown structure for a documentation site	✅ Published (unlisted)

Number	Title	Status
01.09	Copilot settings, IDE support, and compatibility reference	✅ Published

Number	Folder	Topic	Priority	Notes
15.00	04-howto	How to test and iterate on prompts	Medium	Testing strategies, regression detection, A/B comparison
16.00	04-howto	How to version and maintain prompt libraries	Medium	Version control patterns, deprecation, team collaboration
17.00–19.00	04-howto	Reserved	—	Available for future how-to topics

01.00: Series overview and customization stack map
01.01: Appendix (Copilot Spaces)
01.02–01.08: Concept articles (one per customization mechanism)
01.09: Reference article (settings and compatibility)
02.00: Getting started (naming and organizing files)
03.00–09.50: How-to guides — foundations (file types, hooks, models, tools)
10.00–14.00: How-to guides — orchestration and advanced topics (design, subagents, info flow, tokens, SDK)
15.00–19.00: How-to guides — reserved for future topics
20–29: Case studies and applied patterns

Total: 29 articles (28 published + 1 unlisted)

The PE artifact system includes 18 context files, 4 instruction files, 12 agents, 11 prompts, and 2 skills — all documented in the dependency map.

Review	Prompt	Cadence	Next Due
Full system review	`/meta-pe-review`	Biweekly (1st + 15th)	2026-03-15
Optimization pass	`/meta-pe-optimize`	After review (if needed)	As needed
VS Code update check	`/meta-pe-update`	Per VS Code release	Next release

See pe-maintenance.md for the step-by-step guide.

Priority	Item	Status	Notes
Medium	Slim bloated standalone prompts (460–555 lines)	Backlog	Embed PE-validation skill refs instead of inline
Medium	Slim bloated specialist agents (prompt-validator 852 lines)	Backlog	Extract to skill references
Low	Context files 01, 04 over 2,500-token budget	Accepted	Deep-reference docs, loaded on-demand
Low	Evaluate subdirectories for `.github/templates/`	Backlog	Scale readability at 30+ templates

Date	Change	Scope
2026-03-08	Phase 1–6: Full PE artifact improvement plan implemented	All PE artifacts
	— Phase 1: Foundation (dependency map, lifecycle, entry points)	3 new context files
	— Phase 2: Deduplication (~2,570 tokens saved)	01, 08, instructions
	— Phase 3: Meta agents + coherence skill	2 agents, 1 skill
	— Phase 4: Meta prompts for self-improvement	3 meta prompts
	— Phase 5: Renumbered 18 context files into 5-tier logical grouping	All context + refs
	— Phase 6: Maintenance setup (checklist, tasks, roadmap)	Maintenance artifacts

Last updated: 2026-03-08