Virtual Paper Review – Structured Agentic Software Engineering

Join us virtually this Wednesday at 6 pm to continue our monthly Paper Review series!

We’ll unpack how “SE 3.0” organizes human–AI teamwork so outcomes are traceable and reviewable. The paper proposes two coordinated workspaces and a set of lightweight scripts that turn vague tasks into auditable, versioned work.

Part I – Foundations

• Why SE 3.0: from code suggestions to goal-driven engineering
• Two workbenches explained: ACE (people) and AEE (agents)
• Core scripts summarized: BriefingScript, LoopScript, MentorScript
• What “merge-readiness” means as evidence, not opinion

Part II – Deep Dive

• End-to-end flow: ticket → scripts → agent runs → evidence → decision
• Consultation Request Packs (when agents ask for human guidance)
• Merge-Readiness Packs (what evidence reviewers actually need)
• Open research lanes: briefing quality, loop design, tooling, governance, teaming

NOTE – We are moving from Zoom to Google Meet this week for a trial run. I’ll be monitoring this email as we get started up, so please let me know if you have issues joining. You do not need a Google account to join, but I’m sure it will ask to use one if you have it.

Links:

• Main Paper: Agentic Software Engineering: Foundational Pillars and a Research Roadmap — https://arxiv.org/abs/2509.06216

Details:

• Date – 10/08/2025
• Time – 6-7:30pm
• Location – VIRTUAL
• Google Meet – https://meet.google.com/tuq-utxh-riq
• Dial in:‪(US) +1 260-344-4610‬ PIN: ‪207 448 030‬#