1
00:00:00,007 --> 00:00:02,336
I'll show you how I organize my work when I have three or

2
00:00:02,336 --> 00:00:06,068
more projects that are related, or that depend on each other.

3
00:00:09,005 --> 00:00:11,215
On the left are the three projects:

4
00:00:11,215 --> 00:00:14,464
the store API, the portal and payments.

5
00:00:14,464 --> 00:00:18,134
I add one project for each repository, because terminals,

6
00:00:18,134 --> 00:00:21,093
stages and rules are tied to the project.

7
00:00:21,093 --> 00:00:24,887
When I click payments, the Grid and the sidebar switch together.

8
00:00:24,887 --> 00:00:28,296
In Flow, I make one flow for each goal.

9
00:00:28,296 --> 00:00:31,103
This one is for fixing the coupon, and in the selector

10
00:00:31,103 --> 00:00:34,304
there's another one for updating the documentation.

11
00:00:34,304 --> 00:00:37,457
I keep them apart because each one has its own stages,

12
00:00:37,457 --> 00:00:39,902
and I prefer to look at one thing at a time.

13
00:00:39,902 --> 00:00:42,670
To follow several agents, I stay in the Grid,

14
00:00:42,670 --> 00:00:44,892
with the terminals side by side.

15
00:00:44,892 --> 00:00:48,332
When I need to read a long answer, I switch to Focus.

16
00:00:48,332 --> 00:00:50,292
That terminal takes up the center,

17
00:00:50,292 --> 00:00:52,560
and the others stay in the strip at the bottom.

18
00:00:52,560 --> 00:00:55,916
When two agents are going to work in the same repository, I create a

19
00:00:55,916 --> 00:00:57,971
worktree, which is a second folder of

20
00:00:57,971 --> 00:01:01,076
the repository, on a separate branch.

21
00:01:01,076 --> 00:01:03,740
Each agent gets its own copy of the files.

22
00:01:03,740 --> 00:01:08,510
The card shows the line cupom-fix, worktree, one session.

23
00:01:08,510 --> 00:01:11,114
I choose the model based on the task.

24
00:01:11,114 --> 00:01:13,813
I asked Codex for something short, to explain what the

25
00:01:13,813 --> 00:01:17,743
test script is for, and a smaller model handles that.

26
00:01:17,743 --> 00:01:20,036
A bigger model is slower and more expensive,

27
00:01:20,036 --> 00:01:23,445
so I save it for tasks like finding the cause of a bug.

28
00:01:23,445 --> 00:01:27,385
For the stages that matter, I don't rely only on the agent's answer.

29
00:01:27,385 --> 00:01:32,356
The card has the line Verification, npm test, detected.

30
00:01:32,356 --> 00:01:35,535
If a stage has that verification and the tests fail,

31
00:01:35,535 --> 00:01:39,321
it isn't marked done, even if the agent says it's finished.

32
00:01:39,321 --> 00:01:43,091
When I keep repeating a flow, I save it as a routine.

33
00:01:43,091 --> 00:01:48,071
This one, Review and fix, has Claude reviewing and Codex fixing.

34
00:01:48,071 --> 00:01:50,661
And whatever applies to every agent in the project,

35
00:01:50,661 --> 00:01:53,742
like answering in English, I write once in the rules.

36
00:01:53,742 --> 00:01:55,489
In the next one, I'll show what day-to-day

37
00:01:55,489 --> 00:01:58,103
work looks like with more than one project.