Building a Software Factory: Zach Lloyd's Crawl-Walk-Run Framework

X postBen Holmes15 September 2026Resource

Ben Holmes highlights an architectural guide by Zach Lloyd detailing how engineering organisations can progressively adopt an agentic software factory model, moving from isolated point automations to governed cloud workflows.

Key points

Crawl phase: Teams begin with discrete trigger-to-activity tasks such as issue triage, PR code review, monitoring alert responses, or CI self-healing.
Limits of point solutions: Siloed automations lack shared context, create higher maintenance and security surfaces, and prevent end-to-end tracking of metrics like cost-per-PR or cycle time.
Walk phase: Adopting a systemic cloud factory architecture featuring sandboxed cloud environments, centralised agent governance, and comprehensive audit traces.
Human oversight: Cloud factories require reliable escape hatches allowing human engineers to steer running agents or take over tasks locally.
Implementation approach: Whether building custom infrastructure or leveraging platforms like Warp Factories, engineering teams must still build bespoke skills tuned to their specific codebase.

Why it matters

As engineering leaders transition from local interactive coding assistants to autonomous background workflows, ad-hoc setups create operational overhead. A structured maturity framework helps teams scale agentic development securely without fragmenting their delivery pipeline.

Analysis

The analysis underscores the current shift from local agent execution to cloud-native agent loops. While isolated automations using tools like the Claude Code SDK in containers offer quick initial wins, they fail to deliver compounding productivity without a unified context layer. Zach Lloyd frames factory adoption not as a pure build-versus-buy decision, but as foundational platform infrastructure paired with internally developed codebase adaptations.

Original post

Ben Holmes

@BHolmesDev · 15 September 2026

If someone asks you how to set up a software factory, send them this t.co/n6kxOdpiiu

542 likes7 replies1.3k bookmarks124.2k views

View on X

Selected replies (5)

@OMID_0909 @BHolmesDev The factory loop makes a lot of sense. But once multiple agents are working across triage, spec, implementation and verification, shared context becomes more than just memory- you need to know where that knowledge came from, 👇🧵
@spark_dem @BHolmesDev If someone asks you how to set up a software factory, send them this t.co/vsO4BNCxsK
@assaf_elovic @BHolmesDev right framing. the gate to run is review capacity though. if every pr still needs a human yes you just made a faster backlog.
@OMID_0909 @BHolmesDev which version is current, and whether every agent is working from the same source of truth. Do you agree?
@Erkanbalder @BHolmesDev If someone asks you to setup a software factor who doesn’t want push PRs or learn how. Walk away. EOF

Mentioned links

Zach Lloyd on the Crawl, Walk, Run Approach to Software Factoriestwitter.com/zachlloydtweets/status/2099941244063432720

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Similar in content, found by the AI search.

X post:Zach Lloyd15 Oct 2026
The Cultural Shift to Software Factories: Moving from Local Agents to a Central Team Setup
Zach Lloyd shares insights on adopting software factories, noting that the biggest cultural change was transitioning from individual agents running on local machines to a centralised agent setup built collaboratively by the team.
266bookmarks7.3kviews
AI & LLMs· Opinion
Video
Video:CURT6 Oct 2026
Harness Engineering: How to Build a Software Factory with AI Agents
In his talk at the AI Engineer World's Fair, Tessl's Dru Knox introduces the concept of a software factory, where coding agents produce all user-facing code whilst engineers build and maintain the factory infrastructure. He outlines harness engineering (also referred to as loop engineering) as the guiding discipline, dividing engineering work into inner, outer, and meta loops to improve autonomy, automation, and overall code quality.
AI & LLMs· Talk
X post:Zach Lloyd1 Sept 2026
Warp Factories: Automated Skill Improvement via Built-in Feedback Loops
Zach Lloyd shares that Warp's software factory has automatically generated and merged ten skill improvement PRs. This workflow is powered by Warp Factories' new self-improvement feature, which monitors agent runs for inefficiencies and autonomously proposes fixes to agent prompts and skill configurations.
298bookmarks24.3kviews
AI & LLMs· Announcement
X post:Ian Livingstone13 Oct 2026
First Draft of the Agent Baseline Software Factory Reference Architecture Released
Ian Livingstone announced the first draft of the 'Agent Baseline Software Factory Reference Architecture'. Developed alongside community contributors, it offers an opinionated blueprint to secure increasingly autonomous software factories.
263bookmarks16.8kviews
AI & LLMs· Announcement
Link:Warp5 Sept 2026
Warp Factories: Orchestrating Fleets of AI Coding Agents Across the SDLC
Warp has introduced 'Warp Factories' alongside its terminal and CLI tools. The platform provides cloud infrastructure to define, run, and benchmark fleets of autonomous coding agents across the entire software development lifecycle.
AI & LLMs· Announcement
Video
Video:CURT14 Sept 2026
Building a Working Software Factory with AI Agents
Speaking with host Greg, developer Ras Mic details his model-agnostic 'software factory' system. Rather than relying on a commercial platform, the setup consists of five or six markdown files such as agents.md, running an automated four-stage workflow: isolate, build, prove, and ship.
AI & LLMs· Tutorial