Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More
6 contributions to Clief Notes
ICM for Highsfield/ Seedance?
So Im new and still wrapping my head around everything associated with the ICM methodology. One area where I have a lot of questions is how to best use the ICM concept when using video rendering models like Higgsfield/Seedance, Kling, or Veo. I'm wondering if it's possible to connect a Github ICM repository to Higgsfield thru an MCP sever or API Keys. And if I did this would it a) significantly reduce the anount of tokens needed to generate videos, and b) would using an ICM file system like this help with image consistency. I saw somewhere if we created JSON files for images it makes it easier for the image and video softwares to maintain consistency, however that advise was from several months ago; it may be moot by now. If it is helpful, I thought this is what could be in the ICM structure for the videos as well as customer avatar content, etc. Has anyone done something like this, or is this a bad use case for ICM? Thx!
0 likes • 15d
@Ari Evergreen Hiiiiiiiii Ari! Im SO GLAD you're here ! 🤣🤣🤣 I'm prolly gonna need your help at some point. Glad to know we gotta Higgsfield master in the community.
1 like • 15d
@Alessandra Turati Definitely will try with the MCP server if I can.
Mentally Stumped on How to Structure Content
Hello all - Im brand new and have not yet started the training program, although I did read through all of it 1st so that I could conceptually understand what we are building (sorry but that's how my brain works best, 😊). Anyway, I was trying to think about what I need 1st, and it is a Al Agent Coach for my courses. I think this sytem could work well, but I'm stumped on how I would structure the content of my markdown files to make a useful coach. For example - would I break my content down by category of information in each micro section of my teachings, or do I try to anticipate all the questions/problems people have and then build the knowledge base around that? Or do both? Im getting confused because (in my mind), its not the same as what we are learning in the training because its not as structured. For example - theres a specific step to building a video, etc. But a coach has to read the info, then compare it to the student's questions, and then give them tailored responses that align with the concepts in the knowledge base- whether the KB specifically addressed it or not. (This is a spiritual coaching program btw, so many of the answers will be similar in context, but not tge same output). So it's that fluidity Im trying to understand how to best capture. Thanks for any insights that you may have!
2 likes • Jul 21
@Gabriel Azoulay I like this alot. Especially the principles files. Thanks!
3 likes • Jul 21
@Raymond Dennis Omg - this was fantastic thank you! I actually copied everybody's response and put it into claude to give me feedback and the insights were credibly helpful. Your comparison of everyone's feedback also helped me to get clarity about what was really stumping me in terms of approaching this project. The issue is that there are multiple layers of what I'm trying to do that somehow need to be properly organized and captured for this effort to properly work. First, there are the frameworks for the cosmology. There are six total frameworks that interconnect to each other. Additionally, they go fairly in depth to explain all the parts of the frameworks, why they are what they are, and how they relate to everything. So in my mind, this is a level of knowledge base that needs to be created that never changes. The idea is that whatever perspective we're teaching from, we can always reference this content as a type of "doctrine". I'm considering that as one level of information of this knowledge base. I see Aaron's feedback as most closely related to this. Then there needs to be another level of information that is my interpretation of the core content to help the AI agent understand how to interpret the 1st layer of info. This is where Gabriel's feedback really comes in handy. I do want to allow the AI to be able to have some freedom to think, but since I'm speaking on spiritually based content, I dont want the AI to just regurgitate what's being promoted out in the ethos-sphere (internet) because a lot of that is wrong. So I do need to be able to give the AI some direction about how to think about the cosmology content. And I believe this layer of information will help do it. Then there's yet another layer of information that speaks to what are we teaching in the specific courses that the client may be taking as it relates to the previous two layers of information. For me, I'm considering this to be where Carla's feedback is valuable because it helps to tailor the gross level of knowledge base to specific issues.
🛠️ New tool: ICM Architect
I explain the details about it on today's high tea . I built a Claude skill that turns a process, an idea, or a messy folder into an ICM workspace. The folder structure does the orchestration. Numbered folders carry the order, the hierarchy scopes context, and plain markdown files hold state. One agent walks the right files at the right time and does the work a multi-agent setup would. Repo: github.com/RinDig/icm-architect 📦 What it does Two modes. 🔨 Build. You describe your work and it pulls out the structure already sitting in how you talk about it. The stages, the points where you stop and check, what stays the same every run versus what is new. Then it picks one of five proven forms and scaffolds the smallest workspace that carries the job. ♻️ Restructure. Point it at a folder, repo, or vault you already have. It reads every file, sorts each one by role, shows you a migration map, waits for your yes, then moves and checks the result. 🧩 The five forms Pipeline, umbrella, record library, knowledge bundle, context map. They mix and nest, so most real workspaces use more than one. ✅ The walk test Every result gets checked cold. An agent with no memory has to open the root, find its way, act, and report status from the files alone. If it can't, the structure gets fixed until it can. ⚙️ How to use it You can honestly just tell claude to download it from the link, but if you're using codex or something else it will just have to restructure the claude.md to agents.md Or if you want to do more Hands-On install Claude Code: drop the folder in ~/.claude/skills/icm-architect/, then say "ICM this" or "build me a workspace for X." Claude apps: zip the folder and upload it under Settings, then Capabilities. Fork it, break it, tell me what you built. 👇
2 likes • Jul 19
God bless you Jake! I was just wishing this existed! 🤣
🏆Huge Win! Generative AI Ambassador
Hello Clief Notes Family! You may have noticed that I have not been posting as much. Some of that has been intentional, I was posting a lot and being number 1 feels good, but seeing others who work so hard to bring knowledge and value to our community shining feels good also! Another big portion of why I have not been posting is that I have been chasing my dream, I have applied to many ambassador programs and I have been ghosted or I have not been the right fit for them. This does not mean I was not the right fit to be an AI ambassador! I wanted to share, that today is different, today I have won, I have been accepted as Generative AI Ambassador in the company I work for! I will be giving workshops, writing articles and sharing my knowledge with my colleagues everyday, I will be Teaching about AI, 2 of my passions in one place! Teaching and AI. I wanted to thank you all for being one of the most supportive, inspiring, and impactful forces in my life. You have all given me the confidence, the energy, and the care I needed to keep pushing for this. We truly learn, grow, and win together 🤓💪🏆 With all of my heart! THANK YOU 🙏
1 like • Jun 24
🔥🔥🔥 Congratulations! Im glad to hear your tenacity paid off!
Loop of Loops - How are you managing your agent infrastructure
A big part of the ICM method is, and I'm new here so correct any wrong assumptions, is that folders == agent flow. And it allows one agent process to flow naturally into other agent workflows. And I've been thinking about this but I just watched this video by Nate Jones and wanted to get your opinions on it and how it integrates into ICM. https://youtu.be/A4zMyjkL0Dc?is=nPthWUckYYzzIgQY I know the basics map well toba folders workflow structure. But how is this being handled for everyone when multiple sets of workflow folders need to callnornpass information to another one, or even multiple. He hints at one example about planning trips, checking weather, flight info, and more. Some of these loops could easily be called by many other workflows. So when using ICM how are you managing this? Automated calls and starting loops. CRON is an easy answer for answer for basic timed runs. But what about more complex setups. To me this seems like where I would use orchestration managent tools like linear or n8n using recursive workflows. I was hoping to hear how others have solved this while also using the ICM system. Or showing any examples. I've been trying to build some ideas to help manage my own tasks as well and really want to get started with this.
0 likes • Jun 24
I wish I could offer something of value, but I'm confused at the beginning level. Im sure there a way that makes sense & flows w/o a lot of confusion.
1-6 of 6
Yanick Webster
3
8 points to level up
@nikki-webster-6900
Your Expert Next Door: Helps you discover spiritual embodiment with a turnkey on-ramp to financial sovereignty. Embody Your True Divine Self.

Active 2d ago
Joined Jun 18, 2026
INFJ
Powered by