AI tools made developers faster. Nobody prepared the engineering organization for what’s next.

Practical frameworks for building software in the age of AI agents, from the author of The Agentic SDLC Handbook and creator of APM.

Get the Agentic SDLC Handbook

An actionable framework for leaders and practitioners changing how software gets delivered in the agentic era.

Free PDF and EPUB.

You will receive the handbook right away, followed by a short welcome series and The AI Native Mind newsletter (about twice a month). Unsubscribe anytime with one click.

Or read it online for free →

Essay · from the aiDevCon NYC talk, November 2026 Agent Contracts and Receipts: toward verifiable agentic software factories Read the essay → · Format draft 0.1 →

Portrait of Daniel Meppiel

Creator of APM and APMX
Previously GitHub · Sonar · CERN

Connect on LinkedIn

Daniel Meppiel

Director, Agentic SDLC GTM at Microsoft. Author of The Agentic SDLC Handbook.

I believe we are in the early chapters of a fundamental shift in how software gets built: not only faster code generation, but a rethinking of how engineering organizations operate when every developer works alongside AI agents.

I have spent 14+ years on this trajectory: distributed systems at CERN, developer tools at GitHub and Sonar, and founding an AI startup as CTO (WeGaw, backed by the European Space Agency). Today I am a Director at Microsoft, leading GTM for the Agentic SDLC.

I created APM, Microsoft’s open-source Agent Package Manager, now supported by GitHub’s Dependabot and by Renovate. I designed the PROSE specification for writing agent instructions that hold up. I wrote The Agentic SDLC Handbook, everything I have learned about making AI-native development work at scale. And I build APMX, an independent, experimental runner for verifiable agentic software factories.

What you will find in the handbook

  1. Why AI-native development is more than adopting a coding agent. Copilot, Claude or Cursor change how one developer works. What engineering organizations need to change is bigger.
  2. The governance gap. How productivity gains disappear when AI-generated code reaches review, compliance and architecture standards.
  3. Agent onboarding. Patterns that treat coding agents like new hires, with context, standards and guardrails built in.
  4. The PROSE specification. A framework for writing agent instructions that are precise, composable and version-controlled.
  5. A two-command implementation pattern. apm install, then apm compile.
  6. The full method. From one developer’s setup to agent governance across an enterprise.

Get the free handbook Read it online →

Writing

Agent Contracts and Receipts: toward verifiable agentic software factories

Trust in agent-built software can’t come from the model, or from reading its output. A proposal for versioned Agent Contracts, standard receipts and one gate at merge, built on the supply-chain standards software already runs on. Draft for discussion.

November 4, 2026
No matching items

All essays · RSS

Projects

  • APM. Microsoft’s open-source Agent Package Manager: declares the skills, prompts and plugins your coding agents use as dependencies, pins them, and governs them with one org-wide policy.
  • APMX. An independent, experimental open-source runner for agentic software factories: runs Agent Contracts on your coding agent, writes the receipt and provides the gate at merge.
  • Genesis. An architectural layer for designing multi-agent, multi-skill systems with named patterns and contracts, before you write them. It applies the handbook’s architecture patterns to skill design; Appendix B walks through an example.
  • Agent Contracts & Receipts. Format draft 0.1, open for comment: read the spec, join the discussions.