1 00:00:00,000 --> 00:00:06,400 So, I just bumped my local CLI to the new Claude Code, version v2.1.202, 2 00:00:06,500 --> 00:00:10,160 and, uh, they've gone and completely rewritten how code reviews work. 3 00:00:10,293 --> 00:00:13,520 It-it's a massive shift in how the slash commands behave. 4 00:00:13,953 --> 00:00:14,540 Oh, nice. 5 00:00:14,684 --> 00:00:20,620 Well, before we dive into the nuts and bolts of v2.1.202, this quick lightning talk 6 00:00:20,673 --> 00:00:23,100 is brought to you by Jellypod AI. 7 00:00:23,180 --> 00:00:26,540 Now, Lachlan, when you say they rewrote it, are we talking about a complete 8 00:00:26,593 --> 00:00:30,140 deprecation, or did they just, like, split the workflows up? 9 00:00:30,423 --> 00:00:32,343 They split them right down the middle, mate. 10 00:00:32,643 --> 00:00:38,263 See, it used to be that when you typed /review, it would spin up this massive, 11 00:00:38,330 --> 00:00:43,943 multi-agent pipeline that took ages and absolutely chewed through your tokens. 12 00:00:44,123 --> 00:00:49,223 Now, /review is back to being just a fast, single-pass checker. 13 00:00:49,330 --> 00:00:50,503 Super lightweight. 14 00:00:50,796 --> 00:00:56,983 If you want that deep, heavyweight agent pipeline, you have to run /code-review now. 15 00:00:57,170 --> 00:00:59,063 It's its own separate beast. 16 00:00:59,581 --> 00:01:01,128 Ah, that makes so much sense. 17 00:01:01,328 --> 00:01:05,768 I mean, running a full four-agent pipeline every time you just want a quick syntax 18 00:01:05,821 --> 00:01:07,768 check is total overkill. 19 00:01:07,981 --> 00:01:12,888 So, under the hood, is /code-review still doing the whole parallel processing thing? 20 00:01:12,984 --> 00:01:15,848 Like, how does it actually decide if a bug is real? 21 00:01:16,157 --> 00:01:18,157 Yeah, it's pretty wild how they've set it up. 22 00:01:18,317 --> 00:01:20,557 It runs four parallel agents. 23 00:01:20,697 --> 00:01:26,157 You've got two agents strictly checking for compliance with your CLAUDE.md file, 24 00:01:26,217 --> 00:01:30,717 one agent hunting for obvious bugs, and another one digging through your git blame 25 00:01:30,737 --> 00:01:32,397 and history to get context. 26 00:01:32,541 --> 00:01:37,437 Then-and here's the kicker-it has an eighty percent confidence score threshold. 27 00:01:37,570 --> 00:01:41,037 If the agents don't agree with at least eighty percent confidence, 28 00:01:41,090 --> 00:01:42,237 it won't bug you with it. 29 00:01:42,337 --> 00:01:46,877 And they categorize everything into three severities: red circle Important, 30 00:01:46,957 --> 00:01:50,877 yellow circle Nit, and purple circle Pre-existing. 31 00:01:51,064 --> 00:01:55,117 So you know exactly what you actually broke versus what was already dusty. 32 00:01:55,402 --> 00:01:56,762 Wow, eighty percent. 33 00:01:56,906 --> 00:01:58,842 That actually keeps the noise down. 34 00:01:59,002 --> 00:02:02,922 And I love that they call out "Pre-existing" stuff so you don't get blamed for 35 00:02:02,979 --> 00:02:04,122 legacy technical debt. 36 00:02:04,342 --> 00:02:07,562 But, okay, what do we actually do with these findings? 37 00:02:07,695 --> 00:02:11,962 Is it just a report on the screen, or can we, you know, actually execute on them? 38 00:02:12,110 --> 00:02:13,990 Oh, you can absolutely execute. 39 00:02:14,137 --> 00:02:19,630 If you run /code-review --fix, it will actually go ahead and auto-apply those 40 00:02:19,683 --> 00:02:22,430 verified fixes directly to your local files. 41 00:02:22,630 --> 00:02:29,470 But-and please, listen to me on this, fellas-do not run /code-review --fix on 42 00:02:29,523 --> 00:02:30,350 dirty files. 43 00:02:31,184 --> 00:02:35,150 Commit your work first, otherwise you'll be crying into your coffee when it 44 00:02:35,214 --> 00:02:37,310 overwrites your unstaged changes. 45 00:02:38,013 --> 00:02:39,917 Yeah, git commit is your best friend there. 46 00:02:40,029 --> 00:02:43,437 What about CI/CD or PR workflows? 47 00:02:43,477 --> 00:02:45,277 Can we push these comments upstream? 48 00:02:45,583 --> 00:02:46,223 Spot on. 49 00:02:46,303 --> 00:02:51,583 You can run /code-review --comment, and it'll format and post those findings as 50 00:02:51,652 --> 00:02:54,383 inline comments straight onto your GitHub PR. 51 00:02:54,563 --> 00:02:59,343 Now, a quick heads-up on the wallet front: this deep code review is currently a Team 52 00:02:59,383 --> 00:03:03,183 and Enterprise preview, and it does not run on ZDR-you know, 53 00:03:03,279 --> 00:03:05,983 Zero Data Retention-organizations. 54 00:03:06,095 --> 00:03:11,103 Also, if you automate this in your CI pipeline, expect it to cost around fifteen to 55 00:03:11,126 --> 00:03:13,343 twenty-five bucks per run on average. 56 00:03:13,476 --> 00:03:16,863 So, uh, local run control is definitely your friend here. 57 00:03:17,395 --> 00:03:22,142 Oof, twenty-five bucks a run can stack up fast if you have a noisy main branch. 58 00:03:22,251 --> 00:03:24,542 Definitely keep that local first. 59 00:03:24,802 --> 00:03:29,742 Hey, did they do any of those smaller, quality-of-life updates in this version? 60 00:03:29,822 --> 00:03:34,702 Because the previous SSH sign-in flow was a bit of a nightmare with wrapped URLs. 61 00:03:35,005 --> 00:03:35,965 Oh, they fixed that! 62 00:03:36,077 --> 00:03:41,085 They've gone to single-hyperlink sign-in URLs now, so if you're on SSH, 63 00:03:41,138 --> 00:03:43,085 it doesn't wrap and break the link. 64 00:03:43,197 --> 00:03:47,805 They also added OpenTelemetry workflow tags-like workflow.run_id and 65 00:03:47,858 --> 00:03:52,765 workflow.name-so you can actually trace these parallel agents in your APM tools. 66 00:03:52,905 --> 00:03:57,085 And in /config, there's a new "Dynamic workflow size" setting. 67 00:03:57,218 --> 00:04:00,765 It's not a hard cap, just a friendly guideline to tell Claude whether to think 68 00:04:00,832 --> 00:04:02,685 small, medium, or large. 69 00:04:02,872 --> 00:04:07,645 Oh, and if you use voice dictation and it fails to capture your audio a few times, 70 00:04:07,672 --> 00:04:10,765 it now pauses automatically instead of endlessly hanging. 71 00:04:11,250 --> 00:04:14,330 Those trace tags are actually huge for debugging agent loops. 72 00:04:14,440 --> 00:04:21,290 Alright, so the play here is: update to v2.1.202, use /review for your quick 73 00:04:21,347 --> 00:04:26,730 sanity checks, and save /code-review for right before you push-and always, 74 00:04:26,787 --> 00:04:29,450 always commit before running --fix. 75 00:04:29,745 --> 00:04:30,545 Too right, mate. 76 00:04:30,732 --> 00:04:33,985 No one wants to lose half a day's work to an overzealous agent. 77 00:04:34,129 --> 00:04:38,385 Keep it clean, commit early, and let the tools do the heavy lifting. 78 00:04:38,532 --> 00:04:40,145 Catch you on the next one!