"What was my worst lunch?" AI-Powered Blood Sugar Analysis with GitHub Copilot + Claude Code Skills

Hey friends, I'm Scott Hansselman. The Hey friends, I'm Scott Hansselman. The coding revolution, the AI coding coding revolution, the AI coding coding revolution, the AI coding revolution uh is uh in full force right revolution uh is uh in full force right revolution uh is uh in full force right now. It's been going since uh October now. It's been going since uh October now. It's been going since uh October since Opus 4.5 and uh a lot of people since Opus 4.5 and uh a lot of people since Opus 4.5 and uh a lot of people have been talking about skills and it have been talking about skills and it have been talking about skills and it takes a second to get understanding of takes a second to get understanding of takes a second to get understanding of like why would I make a skill? What's a like why would I make a skill? What's a like why would I make a skill? What's a skill do? Is a skill useful? Um, some skill do? Is a skill useful? Um, some skill do? Is a skill useful? Um, some people think skills are just MD files people think skills are just MD files people think skills are just MD files that are just like a markdown skill and that are just like a markdown skill and that are just like a markdown skill and they always start them with things like they always start them with things like they always start them with things like you're an expert. We are almost you're an expert. We are almost you're an expert. We are almost gaslighting the AI agent into thinking gaslighting the AI agent into thinking gaslighting the AI agent into thinking it's good at something that it's not. it's good at something that it's not. it's good at something that it's not. Um, but I've talked about this before. Um, but I've talked about this before. Um, but I've talked about this before. I'm type 1 diabetic. I've got a I'm type 1 diabetic. I've got a I'm type 1 diabetic. I've got a G uh a G7 Dexcom sensor. I've got an G uh a G7 Dexcom sensor. I've got an G uh a G7 Dexcom sensor. I've got an insulin pump. I use an open source insulin pump. I use an open source insulin pump. I use an open source artificial pancreas called Loop that artificial pancreas called Loop that artificial pancreas called Loop that I've been using for almost 10 years. So, I've been using for almost 10 years. So, I've been using for almost 10 years. So, open source makes my uh life and my open source makes my uh life and my open source makes my uh life and my diabetes go around. So, a skill to ask diabetes go around. So, a skill to ask diabetes go around. So, a skill to ask questions about my blood sugar would be questions about my blood sugar would be questions about my blood sugar would be useful. Now, I don't need expertise useful. Now, I don't need expertise useful. Now, I don't need expertise though. I'm not going to try to tell the though. I'm not going to try to tell the though. I'm not going to try to tell the LLM that it's an expert in diabetes. Um, LLM that it's an expert in diabetes. Um, LLM that it's an expert in diabetes. Um, certainly it can go and get PDFs and certainly it can go and get PDFs and certainly it can go and get PDFs and learn about stuff, but I'm an expert in learn about stuff, but I'm an expert in learn about stuff, but I'm an expert in diabetes. I'm an expert in my diabetes.

diabetes. I'm an expert in my diabetes. diabetes. I'm an expert in my diabetes. What I want is one to ask questions What I want is one to ask questions What I want is one to ask questions about like how do Tuesdays go? Uh, how about like how do Tuesdays go? Uh, how about like how do Tuesdays go? Uh, how should I do better when I go to Jimmy should I do better when I go to Jimmy should I do better when I go to Jimmy Johns and have a cookie? Uh, can I Johns and have a cookie? Uh, can I Johns and have a cookie? Uh, can I improve my blood sugar on Thanksgiving? improve my blood sugar on Thanksgiving? improve my blood sugar on Thanksgiving? Stuff like that. So, I made a a Claude Stuff like that. So, I made a a Claude Stuff like that. So, I made a a Claude code skill. Um, I put it into GitHub code skill. Um, I put it into GitHub code skill. Um, I put it into GitHub Copilot CLI. GitHub Copilot CLI is like Copilot CLI. GitHub Copilot CLI is like Copilot CLI. GitHub Copilot CLI is like Cloud Code and it can use Opus, which is Cloud Code and it can use Opus, which is Cloud Code and it can use Opus, which is the 4.5 model that everyone is excited the 4.5 model that everyone is excited the 4.5 model that everyone is excited about. And I'm using a tool called Night about. And I'm using a tool called Night about. And I'm using a tool called Night Scout. Now, NightScout takes the every 5 Scout. Now, NightScout takes the every 5 Scout. Now, NightScout takes the every 5 minutes blood sugar from my Dexcom minutes blood sugar from my Dexcom minutes blood sugar from my Dexcom sensor and uploads it up into a database sensor and uploads it up into a database sensor and uploads it up into a database and it's an open source project and it and it's an open source project and it and it's an open source project and it uses Sometimes it runs in Heroku, uses Sometimes it runs in Heroku, uses Sometimes it runs in Heroku, but I am using a website called T1 PAL. but I am using a website called T1 PAL. but I am using a website called T1 PAL. Um, it's very inexpensive. It's a great Um, it's very inexpensive. It's a great Um, it's very inexpensive. It's a great kind of Kubernetes cluster that you can kind of Kubernetes cluster that you can kind of Kubernetes cluster that you can go and it just handles my night scout. go and it just handles my night scout. go and it just handles my night scout. So, this is my blood sugar right now. So, this is my blood sugar right now. So, this is my blood sugar right now. That is live. It just went down. That's That is live. It just went down. That's That is live. It just went down. That's good. Uh, that's my blood sugar right good. Uh, that's my blood sugar right good. Uh, that's my blood sugar right now. You can see like I had kind of a now. You can see like I had kind of a now. You can see like I had kind of a rough lunch yesterday, but as a general rough lunch yesterday, but as a general rough lunch yesterday, but as a general rule, my blood sugar is pretty much rule, my blood sugar is pretty much rule, my blood sugar is pretty much chilling. Yours is perfect cuz you are chilling. Yours is perfect cuz you are chilling. Yours is perfect cuz you are one of the normally sugared. You're like one of the normally sugared. You're like one of the normally sugared. You're like a you're a muggle, but uh, my blood a you're a muggle, but uh, my blood a you're a muggle, but uh, my blood sugar is all over the place. This messy sugar is all over the place. This messy sugar is all over the place. This messy data here is because I changed my sensor data here is because I changed my sensor data here is because I changed my sensor and the sensors are a little bit sloppy and the sensors are a little bit sloppy and the sensors are a little bit sloppy when they first start up. So, there's a when they first start up. So, there's a when they first start up. So, there's a REST API for that. I can access that REST API for that. I can access that REST API for that. I can access that REST API. So if I go out to the command REST API. So if I go out to the command REST API. So if I go out to the command line here and I see it says 136. Now it line here and I see it says 136. Now it line here and I see it says 136. Now it says 132. I am using uh in this case oh

says 132. I am using uh in this case oh says 132. I am using uh in this case oh my posh uh which is a command line tool my posh uh which is a command line tool my posh uh which is a command line tool that lets you put in a kind of a custom that lets you put in a kind of a custom that lets you put in a kind of a custom prompt. So, we've got the current prompt. So, we've got the current prompt. So, we've got the current folder, my blood sugar, how long the folder, my blood sugar, how long the folder, my blood sugar, how long the last command took, uh the fact that I'm last command took, uh the fact that I'm last command took, uh the fact that I'm on a dev drive, and then, you know, if I on a dev drive, and then, you know, if I on a dev drive, and then, you know, if I did like a get init, uh now I'm on a a did like a get init, uh now I'm on a a did like a get init, uh now I'm on a a branch with an item that's not uh branch with an item that's not uh branch with an item that's not uh currently checked in. So, that's a currently checked in. So, that's a currently checked in. So, that's a prompt. Now, this REST API here lets me prompt. Now, this REST API here lets me prompt. Now, this REST API here lets me see my blood sugar, but it's effectively see my blood sugar, but it's effectively see my blood sugar, but it's effectively a cloud-based open source database of my a cloud-based open source database of my a cloud-based open source database of my blood sugar. Uh so, I thought it'd be blood sugar. Uh so, I thought it'd be blood sugar. Uh so, I thought it'd be cool to make a diabetes skill. Cool. So, cool to make a diabetes skill. Cool. So, cool to make a diabetes skill. Cool. So, I'm going to go and run Copilot. I'm going to go and run Copilot. I'm going to go and run Copilot. Okay. And uh you can install this with Okay. And uh you can install this with Okay. And uh you can install this with Windet. You can install it with npm, Windet. You can install it with npm, Windet. You can install it with npm, with brew. Runs on Linux, runs on Mac, with brew. Runs on Linux, runs on Mac, with brew. Runs on Linux, runs on Mac, runs on Windows. I've got some MCP runs on Windows. I've got some MCP runs on Windows. I've got some MCP servers here, but none of them are servers here, but none of them are servers here, but none of them are diabetes ones. I don't want to make a diabetes ones. I don't want to make a diabetes ones. I don't want to make a diabetes MCP server. An MCP server is diabetes MCP server. An MCP server is diabetes MCP server. An MCP server is USB for AI. It is a collection of tools. USB for AI. It is a collection of tools. USB for AI. It is a collection of tools. Instead, I want to make a skill.

Instead, I want to make a skill. Instead, I want to make a skill. Okay? I'm going to teach it about night Okay? I'm going to teach it about night Okay? I'm going to teach it about night scout. Now, I've done a lot of that work scout. Now, I've done a lot of that work scout. Now, I've done a lot of that work already, but I'm going to show it to already, but I'm going to show it to already, but I'm going to show it to you. So, here's my night scout skill. you. So, here's my night scout skill. you. So, here's my night scout skill. This is up in GitHub. You just clone it This is up in GitHub. You just clone it This is up in GitHub. You just clone it into your skills folder. That easy. It into your skills folder. That easy. It into your skills folder. That easy. It works in claw. It works in um Copilot. works in claw. It works in um Copilot. works in claw. It works in um Copilot. And what it includes is a skill.md that And what it includes is a skill.md that And what it includes is a skill.md that tells the agent about the stuff that it tells the agent about the stuff that it tells the agent about the stuff that it can do, right? And that uh that MD is can do, right? And that uh that MD is can do, right? And that uh that MD is kind of like an instruction manual. It's kind of like an instruction manual. It's kind of like an instruction manual. It's got some PowerShell scripts. This got some PowerShell scripts. This got some PowerShell scripts. This database is local and cached. It's a database is local and cached. It's a database is local and cached. It's a SQLite database because I can bring down SQLite database because I can bring down SQLite database because I can bring down all my blood sugar in a CSV file, but all my blood sugar in a CSV file, but all my blood sugar in a CSV file, but querying a CSV file is a hassle. So, I querying a CSV file is a hassle. So, I querying a CSV file is a hassle. So, I just made a little caching thing here. just made a little caching thing here. just made a little caching thing here. But what's cool about it is that your But what's cool about it is that your But what's cool about it is that your data is your own. It's in NightScout or data is your own. It's in NightScout or data is your own. It's in NightScout or it's local, but you're not giving your it's local, but you're not giving your it's local, but you're not giving your data to an LLM. And and certainly, uh, data to an LLM. And and certainly, uh, data to an LLM. And and certainly, uh, if you wanted to plug this into Olama or if you wanted to plug this into Olama or if you wanted to plug this into Olama or LM Studio and do it entirely local in LM Studio and do it entirely local in LM Studio and do it entirely local in airplane mode, to be clear, airplane mode, to be clear, airplane mode, to be clear, your health data doesn't go anywhere.

your health data doesn't go anywhere. your health data doesn't go anywhere. Now, I've got this scripts folder here Now, I've got this scripts folder here Now, I've got this scripts folder here where I have the cgm py. That's really where I have the cgm py. That's really where I have the cgm py. That's really the skill. So, I'm going to rightclick the skill. So, I'm going to rightclick the skill. So, I'm going to rightclick and I'm going to say open encode. I'm and I'm going to say open encode. I'm and I'm going to say open encode. I'm going to bring that over here. going to bring that over here. going to bring that over here. And let's take a look at that skill And let's take a look at that skill And let's take a look at that skill file. Okay. Provides tools for fetching file. Okay. Provides tools for fetching file. Okay. Provides tools for fetching and analyzing continuous glucose m uh and analyzing continuous glucose m uh and analyzing continuous glucose m uh data from night scout. A cgm that's this data from night scout. A cgm that's this data from night scout. A cgm that's this here. There's lots of different cgms. here. There's lots of different cgms. here. There's lots of different cgms. Night Scout is CGM non-specific. So if Night Scout is CGM non-specific. So if Night Scout is CGM non-specific. So if you have data in Night Scout, if you you have data in Night Scout, if you you have data in Night Scout, if you have a kid or a loved one who has got have a kid or a loved one who has got have a kid or a loved one who has got data in Night Scout, it says data in Night Scout, it says data in Night Scout, it says non-denominational. non-denominational. non-denominational. You just go and um bring this into your You just go and um bring this into your You just go and um bring this into your skills folder and it is a skill that is skills folder and it is a skill that is skills folder and it is a skill that is available. So what I've done now, this available. So what I've done now, this available. So what I've done now, this is just a cgm. by, but it could just as is just a cgm. by, but it could just as is just a cgm. by, but it could just as well be an executable in Rust or well be an executable in Rust or well be an executable in Rust or whatever. But I I use Python because whatever. But I I use Python because whatever. But I I use Python because it's easy and it's good for analyzing it's easy and it's good for analyzing it's easy and it's good for analyzing data, which is also worth pointing out data, which is also worth pointing out data, which is also worth pointing out because I don't want the AI to be doing because I don't want the AI to be doing because I don't want the AI to be doing a lot of math. I want the code to be a lot of math. I want the code to be a lot of math. I want the code to be doing math. LLMs are good at ambiguity.

doing math. LLMs are good at ambiguity. doing math. LLMs are good at ambiguity. They're good at lack of clarity. And They're good at lack of clarity. And They're good at lack of clarity. And where there is a lack of clarity, they where there is a lack of clarity, they where there is a lack of clarity, they strive for clarity, but they use kind of strive for clarity, but they use kind of strive for clarity, but they use kind of the randomization, the next token the randomization, the next token the randomization, the next token prediction of a large language model to prediction of a large language model to prediction of a large language model to go and figure out how to deal with that go and figure out how to deal with that go and figure out how to deal with that ambiguity. But I don't want them using ambiguity. But I don't want them using ambiguity. But I don't want them using ambiguity on my blood sugar. I want to ambiguity on my blood sugar. I want to ambiguity on my blood sugar. I want to do that data with the CGM. So I have do that data with the CGM. So I have do that data with the CGM. So I have like an analyze command and a patterns like an analyze command and a patterns like an analyze command and a patterns command querying and things like that. command querying and things like that. command querying and things like that. But once you give it that tool, like But once you give it that tool, like But once you give it that tool, like when every what do they say? Uh when when every what do they say? Uh when when every what do they say? Uh when everything is a nail, when all you have everything is a nail, when all you have everything is a nail, when all you have is a hammer, everything is a nail. In is a hammer, everything is a nail. In is a hammer, everything is a nail. In this case here, when you're talking this case here, when you're talking this case here, when you're talking about blood sugar, it's going to use about blood sugar, it's going to use about blood sugar, it's going to use this thing. So then I've got a bunch of this thing. So then I've got a bunch of this thing. So then I've got a bunch of instructions about how to use it and instructions about how to use it and instructions about how to use it and then some questions that you could then some questions that you could then some questions that you could potentially ask. What's my blood sugar? potentially ask. What's my blood sugar? potentially ask. What's my blood sugar? Analyze it for the last 30 days. But you Analyze it for the last 30 days. But you Analyze it for the last 30 days. But you can also ask really vague and open-ended can also ask really vague and open-ended can also ask really vague and open-ended questions, which is cool. [snorts] Now, questions, which is cool. [snorts] Now, questions, which is cool. [snorts] Now, that's the preview skill, but I also that's the preview skill, but I also that's the preview skill, but I also wanted to call out the readme, which is wanted to call out the readme, which is wanted to call out the readme, which is for humans. Now, the readme for humans. Now, the readme for humans. Now, the readme also shows that when you're inside of a also shows that when you're inside of a also shows that when you're inside of a of a TUI of a TUI, a text user of a TUI of a TUI, a text user of a TUI of a TUI, a text user interface, uh, a lot of the cool antsy interface, uh, a lot of the cool antsy interface, uh, a lot of the cool antsy and graphics that happen inside of a and graphics that happen inside of a and graphics that happen inside of a TUI, uh, these aren't really graphics, TUI, uh, these aren't really graphics, TUI, uh, these aren't really graphics, you know, that kind of like escape code you know, that kind of like escape code you know, that kind of like escape code graphics get stripped out both in Claude graphics get stripped out both in Claude graphics get stripped out both in Claude and in in uh, GitHub, but you can do and in in uh, GitHub, but you can do and in in uh, GitHub, but you can do heat maps and stuff. So, if you don't heat maps and stuff. So, if you don't heat maps and stuff. So, if you don't like AI and you don't want to do any of like AI and you don't want to do any of like AI and you don't want to do any of this stuff, that's fine. You can still this stuff, that's fine. You can still this stuff, that's fine. You can still take this Python script and do stuff take this Python script and do stuff take this Python script and do stuff with it. Doesn't need to be AI. And you with it. Doesn't need to be AI. And you with it. Doesn't need to be AI. And you can also call these tools from LM Studio can also call these tools from LM Studio can also call these tools from LM Studio and do the work entirely locally like I and do the work entirely locally like I and do the work entirely locally like I said before. So here's how you can just

said before. So here's how you can just said before. So here's how you can just call these scripts directly. I've got a call these scripts directly. I've got a call these scripts directly. I've got a little architecture diagram here. little architecture diagram here. little architecture diagram here. Everything's local. Here's what the AI Everything's local. Here's what the AI Everything's local. Here's what the AI does not have access to. Cool. So I does not have access to. Cool. So I does not have access to. Cool. So I think this is kind of neat. All right, think this is kind of neat. All right, think this is kind of neat. All right, let's see it. So we go in here. let's see it. So we go in here. let's see it. So we go in here. What's my current blood sugar? What's my current blood sugar? What's my current blood sugar? Now [clears throat] I'm using handy uh Now [clears throat] I'm using handy uh Now [clears throat] I'm using handy uh because I don't type well called because I don't type well called because I don't type well called handy.computer which is a local model handy.computer which is a local model handy.computer which is a local model that um you can get it's open source and that um you can get it's open source and that um you can get it's open source and free. You can use dictation you can use free. You can use dictation you can use free. You can use dictation you can use uh whatever makes you happy. Ah this is uh whatever makes you happy. Ah this is uh whatever makes you happy. Ah this is great. So I didn't run the CLI uh I great. So I didn't run the CLI uh I great. So I didn't run the CLI uh I didn't run the the agent in yolo mode. didn't run the the agent in yolo mode. didn't run the the agent in yolo mode. So it has to ask permission every time So it has to ask permission every time So it has to ask permission every time it wants to do something. I could just it wants to do something. I could just it wants to do something. I could just say d- yolo. I could say d-allow all say d- yolo. I could say d-allow all say d- yolo. I could say d-allow all tools. But in this case, I just said, tools. But in this case, I just said, tools. But in this case, I just said, "Let me know if you need anything." So, "Let me know if you need anything." So, "Let me know if you need anything." So, for the very first time, it's like, for the very first time, it's like, for the very first time, it's like, "Well, I really want to call this. It's "Well, I really want to call this. It's "Well, I really want to call this. It's trying to run Python." So, I'm going to trying to run Python." So, I'm going to trying to run Python." So, I'm going to say, "That's cool." Because I know it's say, "That's cool." Because I know it's say, "That's cool." Because I know it's cool. As you get more confident in your cool. As you get more confident in your cool. As you get more confident in your agent, you won't have to do this. Now, agent, you won't have to do this. Now, agent, you won't have to do this. Now, I'm just going to say you can run Python I'm just going to say you can run Python I'm just going to say you can run Python for the rest of our session. You don't for the rest of our session. You don't for the rest of our session. You don't need to keep asking my my questions. And need to keep asking my my questions. And need to keep asking my my questions. And again, I want to point out we're using again, I want to point out we're using again, I want to point out we're using Claude. We're in GitHub Copilot, right, Claude. We're in GitHub Copilot, right, Claude. We're in GitHub Copilot, right, which is a a product that I work on, but which is a a product that I work on, but which is a a product that I work on, but not We are using the Claude comm um the not We are using the Claude comm um the not We are using the Claude comm um the Claude model. Your current blood sugar Claude model. Your current blood sugar Claude model. Your current blood sugar is 128, which is in range from 3 minutes is 128, which is in range from 3 minutes is 128, which is in range from 3 minutes ago. Cool. Um I just had a bagel like an ago. Cool. Um I just had a bagel like an ago. Cool. Um I just had a bagel like an hour and a half ago. Uh what's been hour and a half ago. Uh what's been hour and a half ago. Uh what's been happening over the last hour and a half?

happening over the last hour and a half? happening over the last hour and a half? How's my blood sugar been uh trending How's my blood sugar been uh trending How's my blood sugar been uh trending after that bagel? after that bagel? after that bagel? Now, I'm a bit of a yapper, but I just Now, I'm a bit of a yapper, but I just Now, I'm a bit of a yapper, but I just want to show you that it's that want to show you that it's that want to show you that it's that ambiguity that you can do and ask ambiguity that you can do and ask ambiguity that you can do and ask questions of. So here get you glucose questions of. So here get you glucose questions of. So here get you glucose history for the last 90 minutes. See history for the last 90 minutes. See history for the last 90 minutes. See scripts querying glucose data scripts querying glucose data scripts querying glucose data looking at the options here. Now this is looking at the options here. Now this is looking at the options here. Now this is good. We may have discovered a limit or good. We may have discovered a limit or good. We may have discovered a limit or not a limit rather but you see how it not a limit rather but you see how it not a limit rather but you see how it did it did three queries there. Um I did it did three queries there. Um I did it did three queries there. Um I might need to have hourly instead of might need to have hourly instead of might need to have hourly instead of daily or something like that. Ah it's daily or something like that. Ah it's daily or something like that. Ah it's going to make a spark line chart. It's been flat. That's awesome. Your It's been flat. That's awesome. Your glucose rode rose modestly after the glucose rode rose modestly after the glucose rode rose modestly after the bagel peaked at 166 and it's coming down bagel peaked at 166 and it's coming down bagel peaked at 166 and it's coming down nicely. Solid postmeal response. nicely. Solid postmeal response. nicely. Solid postmeal response. That'sing awesome. Okay, cool. I've been That'sing awesome. Okay, cool. I've been That'sing awesome. Okay, cool. I've been having trouble over lunchtimes. Um the having trouble over lunchtimes. Um the having trouble over lunchtimes. Um the last maybe like two 3 weeks I think I last maybe like two 3 weeks I think I last maybe like two 3 weeks I think I went over 200 several times. um what was went over 200 several times. um what was went over 200 several times. um what was my worst day and what did that day look my worst day and what did that day look my worst day and what did that day look like kind of like in that post lunch like kind of like in that post lunch like kind of like in that post lunch postal postal postal um peak these are the kind of questions diabetes these are the kind of questions diabetes diabetes people ask I've been diabetic diabetes people ask I've been diabetic diabetes people ask I've been diabetic for 30 for 30 for 30 years long time it's not fun I would not years long time it's not fun I would not years long time it's not fun I would not wish this on anybody and um I don't need wish this on anybody and um I don't need wish this on anybody and um I don't need to call a doctor to ask questions like to call a doctor to ask questions like to call a doctor to ask questions like this now I know I could just look at a this now I know I could just look at a this now I know I could just look at a chart and I could put this in Excel all chart and I could put this in Excel all chart and I could put this in Excel all diabetic engineers want to do stuff like diabetic engineers want to do stuff like diabetic engineers want to do stuff like that. But being able to ask a question

that. But being able to ask a question that. But being able to ask a question of an agent like this and have it be of an agent like this and have it be of an agent like this and have it be totally um local code. Everything's totally um local code. Everything's totally um local code. Everything's happening locally here um is pretty happening locally here um is pretty happening locally here um is pretty cool. cool. cool. Okay. So look, there's the the SQL stuff Okay. So look, there's the the SQL stuff Okay. So look, there's the the SQL stuff there. Oh, this is interesting. Okay, there. Oh, this is interesting. Okay, there. Oh, this is interesting. Okay, this is great. So let's talk about this this is great. So let's talk about this this is great. So let's talk about this as developers. So we're going to go into as developers. So we're going to go into as developers. So we're going to go into developer mind here. In this case here, developer mind here. In this case here, developer mind here. In this case here, it's looking at the code. It's trying to it's looking at the code. It's trying to it's looking at the code. It's trying to look at the patterns from my code, look at the patterns from my code, look at the patterns from my code, right? It look at the last 3 weeks right? It look at the last 3 weeks right? It look at the last 3 weeks starting at 11 and ending at 3:00. And starting at 11 and ending at 3:00. And starting at 11 and ending at 3:00. And now it's querying the database directly now it's querying the database directly now it's querying the database directly because it didn't find something it because it didn't find something it because it didn't find something it needed in my tool. So I'm actually going needed in my tool. So I'm actually going needed in my tool. So I'm actually going to be able to improve my tool here. to be able to improve my tool here. to be able to improve my tool here. Okay. Okay. Okay. So here's our database schema. It's So here's our database schema. It's So here's our database schema. It's going to look at that table. Let's query the database quer. Uh so Let's query the database quer. Uh so it's quer it's making a database query it's quer it's making a database query it's quer it's making a database query on its own. Again, local. You can see on its own. Again, local. You can see on its own. Again, local. You can see that local database. Some queries are fast like you saw it Some queries are fast like you saw it asked my blood sugar instantly. But this asked my blood sugar instantly. But this asked my blood sugar instantly. But this was a sophisticated and complex query.

was a sophisticated and complex query. was a sophisticated and complex query. Find worst lunch. It's making a Find worst lunch. It's making a Find worst lunch. It's making a temporary folder here, a temporary file temporary folder here, a temporary file temporary folder here, a temporary file here, here, here, and it's creating a get the worst day. and it's creating a get the worst day. and it's creating a get the worst day. That's fun. It's putting it in a temp That's fun. It's putting it in a temp That's fun. It's putting it in a temp folder here called session state to see folder here called session state to see folder here called session state to see if that works out. You went over 200 on 10 different days. You went over 200 on 10 different days. Your worst day was yesterday. Yeah, Your worst day was yesterday. Yeah, Your worst day was yesterday. Yeah, yesterday sucked. yesterday sucked. yesterday sucked. Shows the timeline. Big spike. 200 Shows the timeline. Big spike. 200 Shows the timeline. Big spike. 200 points. 200 points from trough to peak points. 200 points from trough to peak points. 200 points from trough to peak in two hours. Other bad days. Okay, so in two hours. Other bad days. Okay, so in two hours. Other bad days. Okay, so now here's programmer mode. I'm going to now here's programmer mode. I'm going to now here's programmer mode. I'm going to take my diabetic hat off with my take my diabetic hat off with my take my diabetic hat off with my programmer hat on. I noticed that you programmer hat on. I noticed that you programmer hat on. I noticed that you had to do a custom query for that. And had to do a custom query for that. And had to do a custom query for that. And um this might be a common thing, but I um this might be a common thing, but I um this might be a common thing, but I think find worst lunch. py is not the think find worst lunch. py is not the think find worst lunch. py is not the answer. What is the generalized problem answer. What is the generalized problem answer. What is the generalized problem that you needed to write a custom uh that you needed to write a custom uh that you needed to write a custom uh Python file for and how can we improve Python file for and how can we improve Python file for and how can we improve the skill so that you could do things the skill so that you could do things the skill so that you could do things like this later? I think people might like this later? I think people might like this later? I think people might ask about dinner, about lunch, about ask about dinner, about lunch, about ask about dinner, about lunch, about breakfast. Um, this query didn't seem breakfast. Um, this query didn't seem breakfast. Um, this query didn't seem like you needed to write a specific like you needed to write a specific like you needed to write a specific skill. How can we improve the the main skill. How can we improve the the main skill. How can we improve the the main Python uh cgm py to do stuff like this Python uh cgm py to do stuff like this Python uh cgm py to do stuff like this in the future? uh suggest a couple of in the future? uh suggest a couple of in the future? uh suggest a couple of things before you make changes.

things before you make changes. things before you make changes. Now, you might laugh at me. You might Now, you might laugh at me. You might Now, you might laugh at me. You might think that, wow, this is like, you know, think that, wow, this is like, you know, think that, wow, this is like, you know, professional engineer and he's yapping professional engineer and he's yapping professional engineer and he's yapping like this, but it's a rubber duck, like this, but it's a rubber duck, like this, but it's a rubber duck, right? You got to where's my duck? right? You got to where's my duck? right? You got to where's my duck? Right? Rubber ducking is calling your Right? Rubber ducking is calling your Right? Rubber ducking is calling your buddy and like where you put the rubber buddy and like where you put the rubber buddy and like where you put the rubber duck on your on your uh monitor and you duck on your on your uh monitor and you duck on your on your uh monitor and you talk to it. You talk it through. That's talk to it. You talk it through. That's talk to it. You talk it through. That's too vague for it to probably, you know, too vague for it to probably, you know, too vague for it to probably, you know, fix for me, but I can brainstorm with fix for me, but I can brainstorm with fix for me, but I can brainstorm with it. Right? So in this case here, I'm it. Right? So in this case here, I'm it. Right? So in this case here, I'm brainstorming. So it's suggesting a day brainstorming. So it's suggesting a day brainstorming. So it's suggesting a day command, a worst command, extreme days, command, a worst command, extreme days, command, a worst command, extreme days, and meal presets. Yeah, I like A and B. and meal presets. Yeah, I like A and B. and meal presets. Yeah, I like A and B. Day and worst are a good idea. Let's do Day and worst are a good idea. Let's do Day and worst are a good idea. Let's do A and B. Day is a miscalculation on my part. I Day is a miscalculation on my part. I should have done that. I think I did should have done that. I think I did should have done that. I think I did month and week. Um, month and week. Um, month and week. Um, meal presets, you know, if you're meal presets, you know, if you're meal presets, you know, if you're Spanish, eat dinner at night. having mid Spanish, eat dinner at night. having mid Spanish, eat dinner at night. having mid dinner presets don't make any sense dinner presets don't make any sense dinner presets don't make any sense right now, but this is open source, right now, but this is open source, right now, but this is open source, right? You can do anything you want to, right? You can do anything you want to, right? You can do anything you want to, which is cool. Uh, yeah, covers the find which is cool. Uh, yeah, covers the find which is cool. Uh, yeah, covers the find a problem drill into it that I just did a problem drill into it that I just did a problem drill into it that I just did manually. So, here we go. Now, I can go manually. So, here we go. Now, I can go manually. So, here we go. Now, I can go back over into Visual Studio Code and I back over into Visual Studio Code and I back over into Visual Studio Code and I can watch this work that's happening can watch this work that's happening can watch this work that's happening because this is just making changes to because this is just making changes to because this is just making changes to that script. You can see the script's that script. You can see the script's that script. You can see the script's been modified right there.

been modified right there. been modified right there. All right. And we're making a day All right. And we're making a day All right. And we're making a day change. There you go. Analyze days. change. There you go. Analyze days. change. There you go. Analyze days. Now my data my my my night scout URL is Now my data my my my night scout URL is Now my data my my my night scout URL is staying in an environment variable. I staying in an environment variable. I staying in an environment variable. I just set that in an environment just set that in an environment just set that in an environment variable. Now we're going to add two variable. Now we're going to add two variable. Now we're going to add two functions before main. functions before main. functions before main. This is good. Now if I want to see This is good. Now if I want to see This is good. Now if I want to see what's happening, see where it says what's happening, see where it says what's happening, see where it says controlr? I can go and hit control R and controlr? I can go and hit control R and controlr? I can go and hit control R and just watch it watch it work and I can just watch it watch it work and I can just watch it watch it work and I can steer it. So steering means talking to steer it. So steering means talking to steer it. So steering means talking to the AI while it's working. It might seem the AI while it's working. It might seem the AI while it's working. It might seem as you're watching this video that this as you're watching this video that this as you're watching this video that this is one of those things where I just yap is one of those things where I just yap is one of those things where I just yap and this thing writes code, but you'll and this thing writes code, but you'll and this thing writes code, but you'll find as a professional programmer, find as a professional programmer, find as a professional programmer, you're going to spend about half your you're going to spend about half your you're going to spend about half your time designing and talking and half your time designing and talking and half your time designing and talking and half your time reviewing code. Uh this is not vibe time reviewing code. Uh this is not vibe time reviewing code. Uh this is not vibe coding. This is like AI augmented coding. This is like AI augmented coding. This is like AI augmented coding. In this case here, I'm going to coding. In this case here, I'm going to coding. In this case here, I'm going to look at the code, think about the code, look at the code, think about the code, look at the code, think about the code, and uh they've added slashmeal. Let me and uh they've added slashmeal. Let me and uh they've added slashmeal. Let me go look at that real quick. So it's go look at that real quick. So it's go look at that real quick. So it's working while I'm working. Here's the working while I'm working. Here's the working while I'm working. Here's the meal. I don't really want the presets. I meal. I don't really want the presets. I meal. I don't really want the presets. I think it put meal presets in there think it put meal presets in there think it put meal presets in there anyway. So, that's a thing. I'll have to anyway. So, that's a thing. I'll have to anyway. So, that's a thing. I'll have to think about that. I don't like think about that. I don't like think about that. I don't like hard-coded stuff. There's the fetch and hard-coded stuff. There's the fetch and hard-coded stuff. There's the fetch and store.

It did worst. Anyway, It did worst. Anyway, did I ask it to do worst? Oh, I did do I did I ask it to do worst? Oh, I did do I did I ask it to do worst? Oh, I did do I asked it to do worst. Why do we have asked it to do worst. Why do we have asked it to do worst. Why do we have meal presets? meal presets? meal presets? That's cool. Um, That's cool. Um, That's cool. Um, why do we do mail presets? Did we why do we do mail presets? Did we why do we do mail presets? Did we conflate A, B, and C or where is that conflate A, B, and C or where is that conflate A, B, and C or where is that hardcoded? I mean, I guess there's no hardcoded? I mean, I guess there's no hardcoded? I mean, I guess there's no way to do that without way to do that without way to do that without Okay, let's think about this. Yeah, I feel like C got conflated into B Yeah, I feel like C got conflated into B because if I go back over here and we because if I go back over here and we because if I go back over here and we look at worst. Yeah. Best worst hours. Best worth days, Yeah. Best worst hours. Best worth days, problematic day hour combos. problematic day hour combos. problematic day hour combos. Yeah, that seems clean. Yeah, that seems clean. Yeah, that seems clean. That's That's the pattern thing I did That's That's the pattern thing I did That's That's the pattern thing I did earlier. earlier. earlier. So, where is worst now?

Find worst days. Find worst days. Yep, there we go. Peak trough glucose. Yep, there we go. Peak trough glucose. Yep, there we go. Peak trough glucose. SGV stands for serum glucose value. All SGV stands for serum glucose value. All SGV stands for serum glucose value. All of this is being done locally in the of this is being done locally in the of this is being done locally in the database. That's one of the things nice database. That's one of the things nice database. That's one of the things nice about it. You're right to question it. about it. You're right to question it. about it. You're right to question it. Meal presets are hardcoded. Everyone's Meal presets are hardcoded. Everyone's Meal presets are hardcoded. Everyone's meal times are configurable. Yeah, I don't want hardcoded meal Yeah, I don't want hardcoded meal presets. Yeah, I would prefer not to presets. Yeah, I would prefer not to presets. Yeah, I would prefer not to have meal presets that are um have meal presets that are um have meal presets that are um hardcoded. Everyone eats lunch and meals hardcoded. Everyone eats lunch and meals hardcoded. Everyone eats lunch and meals at different times. And this needs to at different times. And this needs to at different times. And this needs to work in other languages and other work in other languages and other work in other languages and other countries. Let's just do hour start and countries. Let's just do hour start and countries. Let's just do hour start and hour end. Let's get rid of any hour end. Let's get rid of any hour end. Let's get rid of any hard-coded presets. And if you're and hard-coded presets. And if you're and hard-coded presets. And if you're and you can put in the skill if someone asks you can put in the skill if someone asks you can put in the skill if someone asks a question about a meal, then you can a question about a meal, then you can a question about a meal, then you can ask them, hey, when is a time uh that I ask them, hey, when is a time uh that I ask them, hey, when is a time uh that I start my meals? And then do your hour start my meals? And then do your hour start my meals? And then do your hour and uh hour end from there. So this will and uh hour end from there. So this will and uh hour end from there. So this will mean that the skill will get a little mean that the skill will get a little mean that the skill will get a little bit smarter. it will pull out that that bit smarter. it will pull out that that bit smarter. it will pull out that that bit of code there that was going to have bit of code there that was going to have bit of code there that was going to have hard-coded meals. I don't want anything hard-coded meals. I don't want anything hard-coded meals. I don't want anything hard-coded in here because this is the hard-coded in here because this is the hard-coded in here because this is the thing that's so important about thing that's so important about thing that's so important about diabetes. Everybody's diabetes is diabetes. Everybody's diabetes is diabetes. Everybody's diabetes is different. When I said before kind of different. When I said before kind of different. When I said before kind of half jokingly like I'm an expert in half jokingly like I'm an expert in half jokingly like I'm an expert in diabetes. I'm an expert in mine. I know diabetes. I'm an expert in mine. I know diabetes. I'm an expert in mine. I know more about my diabetes than my doctor more about my diabetes than my doctor more about my diabetes than my doctor does. So, you never want when you're does. So, you never want when you're does. So, you never want when you're building diabetes tools like this to building diabetes tools like this to building diabetes tools like this to assume anything about someone else's um assume anything about someone else's um assume anything about someone else's um blood sugar. Super important. So, we'll blood sugar. Super important. So, we'll blood sugar. Super important. So, we'll get rid of that meal stuff and we'll get rid of that meal stuff and we'll get rid of that meal stuff and we'll clean this up. And then now, if I ask it clean this up. And then now, if I ask it clean this up. And then now, if I ask it about lunch, it'll say, "Well, when

about lunch, it'll say, "Well, when about lunch, it'll say, "Well, when about you eat lunch?" And I do it from about you eat lunch?" And I do it from about you eat lunch?" And I do it from like noon to 2. So, we'll just kind of like noon to 2. So, we'll just kind of like noon to 2. So, we'll just kind of yank that out of there. yank that out of there. yank that out of there. And um and notice that we're just doing And um and notice that we're just doing And um and notice that we're just doing it in one cg. py now because I'm kind of it in one cg. py now because I'm kind of it in one cg. py now because I'm kind of making not a not a MCP server, but a making not a not a MCP server, but a making not a not a MCP server, but a multipplexer, kind of like a a a cgm. Is multipplexer, kind of like a a a cgm. Is multipplexer, kind of like a a a cgm. Is a pile of tools. Um, and it's just got a pile of tools. Um, and it's just got a pile of tools. Um, and it's just got you call it calls cgm.c py when it wants you call it calls cgm.c py when it wants you call it calls cgm.c py when it wants to talk to the cgm and then it's got d- to talk to the cgm and then it's got d- to talk to the cgm and then it's got d- this that. There's like five or 10 of this that. There's like five or 10 of this that. There's like five or 10 of these. I want to keep it clean. Um, and these. I want to keep it clean. Um, and these. I want to keep it clean. Um, and I wanted to just talk to that local I wanted to just talk to that local I wanted to just talk to that local database for speed. Um, certainly I database for speed. Um, certainly I database for speed. Um, certainly I could have it call night scout, but it could have it call night scout, but it could have it call night scout, but it only needs to call night scout when it only needs to call night scout when it only needs to call night scout when it refreshes refreshes the data. refreshes refreshes the data. refreshes refreshes the data. Okay, clean up that and get rid of the Okay, clean up that and get rid of the Okay, clean up that and get rid of the meal parameter. probably going to start meal parameter. probably going to start meal parameter. probably going to start to have to write unit tests around this to have to write unit tests around this to have to write unit tests around this because I have my spec in my readme. Um, because I have my spec in my readme. Um, because I have my spec in my readme. Um, but I'll need to get more formal if this but I'll need to get more formal if this but I'll need to get more formal if this thing gets too complicated. There we go. thing gets too complicated. There we go. thing gets too complicated. There we go. There's my worst. See, now it's doing There's my worst. See, now it's doing There's my worst. See, now it's doing hour start, hour end. That's what I hour start, hour end. That's what I hour start, hour end. That's what I want. Worst command without meal preset.

want. Worst command without meal preset. want. Worst command without meal preset. Perfect. Perfect. Perfect. Okay, cool. What was my worst lunch this Okay, cool. What was my worst lunch this Okay, cool. What was my worst lunch this last week? I really like handy. Handy.comput. I really like handy. Handy.comput. That's the website. What? Okay. I That's the website. What? Okay. I That's the website. What? Okay. I typically start lunch between 11 and I'm typically start lunch between 11 and I'm typically start lunch between 11 and I'm done by two. So somewhere in there. So now I would expect it to There it is. So now I would expect it to There it is. Hour start, hour end. That's useful. Hour start, hour end. That's useful. Hour start, hour end. That's useful. Perfect. Fantastic. Uh yeah, give me a Perfect. Fantastic. Uh yeah, give me a Perfect. Fantastic. Uh yeah, give me a timeline, but actually do it as a um a timeline, but actually do it as a um a timeline, but actually do it as a um a spark line uh chart and then also give spark line uh chart and then also give spark line uh chart and then also give me the command so I can see that in full me the command so I can see that in full me the command so I can see that in full color in another terminal. color in another terminal. color in another terminal. So, some of this will use color and some So, some of this will use color and some So, some of this will use color and some of it won't. But when you're inside of a of it won't. But when you're inside of a of it won't. But when you're inside of a of a TUI like this, sometimes of a TUI like this, sometimes of a TUI like this, sometimes uh those antsy bits get uh uh those antsy bits get uh uh those antsy bits get uh those antsy bits get stripped out. So, those antsy bits get stripped out. So, those antsy bits get stripped out. So, I'm going to actually have it give me I'm going to actually have it give me I'm going to actually have it give me the command because I wanted to show you the command because I wanted to show you the command because I wanted to show you also that you don't need to do any of also that you don't need to do any of also that you don't need to do any of this this way. There you go. There's my this this way. There you go. There's my this this way. There you go. There's my there's my peak there. Okay. So now this there's my peak there. Okay. So now this there's my peak there. Okay. So now this here is the command I can run elsewhere.

here is the command I can run elsewhere. here is the command I can run elsewhere. Okay. Oh, I'm going to add a date. Okay. Oh, I'm going to add a date. Okay. Oh, I'm going to add a date. What's the difference between day and What's the difference between day and What's the difference between day and date? date? date? Oh, that'd be cool. Sure, do that. Oh, that'd be cool. Sure, do that. Oh, that'd be cool. Sure, do that. While it's doing that, I'm going to run While it's doing that, I'm going to run While it's doing that, I'm going to run this real quick. Oops. Oops. I need to I need to I need to copy paste that script. Python. This might break because it Python. This might break because it might be working on the code at the same might be working on the code at the same might be working on the code at the same time. time. time. Yeah, that's what that looks like. But I Yeah, that's what that looks like. But I Yeah, that's what that looks like. But I want a spark line. want a spark line. want a spark line. Cool. So, we're going to get a spark Cool. So, we're going to get a spark Cool. So, we're going to get a spark line. And spark lines are great and they line. And spark lines are great and they line. And spark lines are great and they look really cool in um the command line. look really cool in um the command line. look really cool in um the command line. So I have a chart command. Now I'm not So I have a chart command. Now I'm not So I have a chart command. Now I'm not making charts that are like graphical making charts that are like graphical making charts that are like graphical charts. I'm doing it with um Anie. So charts. I'm doing it with um Anie. So charts. I'm doing it with um Anie. So I've got spark lines. I've got heat I've got spark lines. I've got heat I've got spark lines. I've got heat maps. Got all kinds of stuff. Actually maps. Got all kinds of stuff. Actually maps. Got all kinds of stuff. Actually I'll show you what that looks like. So I'll show you what that looks like. So I'll show you what that looks like. So if we go over to let me bring that up on if we go over to let me bring that up on if we go over to let me bring that up on another another deal here.

we go. we go. So, here's what like my heat map looks So, here's what like my heat map looks So, here's what like my heat map looks like, which is cool. So, I've got other like, which is cool. So, I've got other like, which is cool. So, I've got other like I heard it just pop up a little like I heard it just pop up a little like I heard it just pop up a little noise there noise there noise there telling me that it was done. All right. telling me that it was done. All right. telling me that it was done. All right. So, I'm going to run this So, I'm going to run this So, I'm going to run this in another terminal. Why is it in there? Twist. Am I doing Why is it in there? Twist. Am I doing something wrong? something wrong? something wrong? Python.

There we go. Yeah. See? So now I got a There we go. Yeah. See? So now I got a nice spark line. I can do bigger charge. nice spark line. I can do bigger charge. nice spark line. I can do bigger charge. You can do whatever you want, right? But You can do whatever you want, right? But You can do whatever you want, right? But I can see that like that that that 4hour I can see that like that that that 4hour I can see that like that that that 4hour period was not amazing and my time in period was not amazing and my time in period was not amazing and my time in range was not awesome. But that that's range was not awesome. But that that's range was not awesome. But that that's still pretty useful. Black and white still pretty useful. Black and white still pretty useful. Black and white over here. [snorts] Cool. This looks over here. [snorts] Cool. This looks over here. [snorts] Cool. This looks great. Make sure the skills file and great. Make sure the skills file and great. Make sure the skills file and read me file both reflect this new read me file both reflect this new read me file both reflect this new reality. And check this in. reality. And check this in. reality. And check this in. Oh, and make a really good commit Oh, and make a really good commit Oh, and make a really good commit message. message. message. Okay. Okay. Okay. So, now we'll switch back over to uh So, now we'll switch back over to uh So, now we'll switch back over to uh GitHub. Oop, it's asking me a question. Ah, so I'm, this is a little Ah, so I'm, this is a little interesting. I'm doing the changes interesting. I'm doing the changes interesting. I'm doing the changes directly in the skills folder. So I directly in the skills folder. So I directly in the skills folder. So I don't have like a working folder. I'm don't have like a working folder. I'm don't have like a working folder. I'm editing the skill while I'm using the editing the skill while I'm using the editing the skill while I'm using the skill, which is kind of cool. And then skill, which is kind of cool. And then skill, which is kind of cool. And then it's going to use the GitHub GH CLI to it's going to use the GitHub GH CLI to it's going to use the GitHub GH CLI to just post that up into uh there it is. just post that up into uh there it is. just post that up into uh there it is. It's going to update both files.

It's going to update both files. It's going to update both files. And it's going up here. And it's going up here. And it's going up here. And then we're going to go under commits And then we're going to go under commits And then we're going to go under commits here. You see the last commit was 37 here. You see the last commit was 37 here. You see the last commit was 37 minutes ago. Updating the docs. minutes ago. Updating the docs. minutes ago. Updating the docs. And then in a moment we will see an And then in a moment we will see an And then in a moment we will see an update. Right now it says 40 minutes update. Right now it says 40 minutes update. Right now it says 40 minutes ago. ago. ago. In a second it's going to say 10 seconds In a second it's going to say 10 seconds In a second it's going to say 10 seconds ago. Okay. And then we'll see the readme update as And then we'll see the readme update as well. well. well. All right. There we go. Updating the All right. There we go. Updating the All right. There we go. Updating the readme. updating the skill. I can take a readme. updating the skill. I can take a readme. updating the skill. I can take a look at those skills. Take a look at look at those skills. Take a look at look at those skills. Take a look at those changes. I can go in here and look those changes. I can go in here and look those changes. I can go in here and look at the diffs at the diffs at the diffs and see what's changed. Yeah, there's and see what's changed. Yeah, there's and see what's changed. Yeah, there's our worst day finder and our day our worst day finder and our day our worst day finder and our day analysis that's rolling in right now. analysis that's rolling in right now. analysis that's rolling in right now. Okay, let's just modify these lines of Okay, let's just modify these lines of Okay, let's just modify these lines of the readme. Now, I'll need to go through the readme. Now, I'll need to go through the readme. Now, I'll need to go through all of this tonight and make sure that all of this tonight and make sure that all of this tonight and make sure that it's all clean because I'm doing this it's all clean because I'm doing this it's all clean because I'm doing this for the purposes of the video, but I for the purposes of the video, but I for the purposes of the video, but I just want to show you how easy it is to just want to show you how easy it is to just want to show you how easy it is to make a skill. There's worst and there's make a skill. There's worst and there's make a skill. There's worst and there's now it wants to know if it can go and do now it wants to know if it can go and do now it wants to know if it can go and do this. We're going to go ahead and check this. We're going to go ahead and check this. We're going to go ahead and check our git status, update the docs and our git status, update the docs and our git status, update the docs and commit. Notice again, it's asking commit. Notice again, it's asking commit. Notice again, it's asking permission at every at every step. Any permission at every at every step. Any permission at every at every step. Any good coding agent, whether it be cloud good coding agent, whether it be cloud good coding agent, whether it be cloud code or GitHub, does that. There's our code or GitHub, does that. There's our code or GitHub, does that. There's our git add and our git commit. It's going git add and our git commit. It's going git add and our git commit. It's going to put in a nice long commit message.

Okay, let's go back over and take a Okay, let's go back over and take a look. Oh, did it not push? Oh, did it not push? Probably didn't push. Okay, push. I Probably didn't push. Okay, push. I Probably didn't push. Okay, push. I could get push if I needed to, but could get push if I needed to, but could get push if I needed to, but getting lazy now. There it is. Yes, do getting lazy now. There it is. Yes, do getting lazy now. There it is. Yes, do the push. There's the push now. Now, now. There's the push now. Now, now. Look at that. Scrumptious. Look at that. Scrumptious. Look at that. Scrumptious. Cool. So, that is uh making a skill, a Cool. So, that is uh making a skill, a Cool. So, that is uh making a skill, a night scout skill in this case for type night scout skill in this case for type night scout skill in this case for type 1 diabetes with GitHub Copilot CLI. Uh 1 diabetes with GitHub Copilot CLI. Uh 1 diabetes with GitHub Copilot CLI. Uh let me know if you want me to do let me know if you want me to do let me know if you want me to do anything else in the comments. Thanks so anything else in the comments. Thanks so anything else in the comments. Thanks so much for watching.