Agentic Software Factory workshop · July 17, 2026

Navigate the recording by outcome

Choose a chapter and the embedded recording cues its exact start. Every chapter URL can be bookmarked, copied, and restored.

Full workshop · 05:53:18

The complete recording

The recording stays unloaded on a plain visit. Starting or following a chapter link loads only the media needed to cue that moment.

Workshop video

Choose a chapter below to cue the video to its exact start.

Choose a chapter or press Play.

Open the full recording in Dropbox

01

00:15:52–00:50:18

Bound one factory loop

Turn one recurring task into an inspectable loop with inputs, permissions, proof, a human gate, and a durable receipt.

“If the agents can't see it, it doesn't exist.”

John Lindquist · 00:23:25
Watch chapterOpen lesson
02

00:57:58–01:16:45

Put policy at the narrowest boundary

Make stable intent and risk rules arrive before the action they should change.

“Meaning that… It won't ever miss because you're configuring the place where the agent is running.”

John Lindquist · 01:02:42
Watch chapterOpen lesson
03

01:21:23–02:01:26

Keep wrappers thin and work visible

Remove flag memorization while keeping the real command, task state, logs, and verification inspectable.

“Kind of forget the flags, especially if you're someone jumping back and forth between the different harnesses.”

John Lindquist · 01:23:01
Watch chapterOpen lesson
04

02:15:42–02:34:33

Point at the visual problem

Replace vague design prose with an annotated image, a bounded target, and a visible success condition.

“Screenshot, draw an arrow, or multiple arrows, and then paste.”

John Lindquist · 02:18:19
Watch chapterOpen lesson
05

02:34:33–02:51:53

Compose only the evidence the task needs

Build a prompt packet from text, files, images, and fresh state without smuggling in an entire stale session.

“How can I consolidate that into something that's gonna manage it for me?”

John Lindquist · 02:37:56
Watch chapterOpen lesson
06

02:51:53–03:38:37

Turn good prompts into inspectable flows

Reuse a task recipe that can inject fresh command output, preserve prompt structure, and expose what the agent will receive.

“If you think of managing context as having only the context that I need for the specific task that I want.”

John Lindquist · 02:55:29
Watch chapterOpen lesson
07

04:20:03–04:42:47

Recover state before continuing

Reconstruct the smallest trustworthy state packet instead of asking a new session to absorb stale scrollback.

“Just ask the agent, or ask yourself: how could I break this into a smaller task?”

John Lindquist · 04:28:49
Watch chapterOpen lesson
08

04:42:47–05:05:27

Diverge first, then record taste

Make important design choices visible while change is cheap, then turn preference into a reusable contract.

“When code is disposable, code is cheap.”

John Lindquist · 04:43:09
Watch chapterOpen lesson
09

05:05:27–05:25:50

Escalate the decision, not the workflow

Spend frontier-model or video-analysis overhead only where a second reasoning or observation path can change the decision.

“Might as well use the maximum intelligence I have access to.”

John Lindquist · 05:11:10
Watch chapterOpen lesson
10

05:25:50–05:40:36

Let vision guide automation

Keep code as implementation truth while giving agents durable product direction and an explicit adherence dial.

“Code is a source of truth.”

John Lindquist · 05:35:33
Watch chapterOpen lesson
11

05:40:36–05:53:13

Close the loop without hiding open work

Separate answered questions, dated claims, standing offers, and still-open deliverables.

“And how important to you is it that it's 100% correct or even better than you expected is an important factor.”

John Lindquist · 05:44:28
Watch chapterOpen lesson

Prepared after the recording

12

Prepared follow-up · no dedicated recording section

Attach testimony to runtime evidence

Give a debugging agent the exact moment of failure, the human-observed symptom, and a bounded event trail.

No dedicated recording section

Open lesson