Activity
Mon
Wed
Fri
Sun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
What is this?
Less
More
Memberships
Clief Notes
37.3k members • Free
112 contributions to Clief Notes
6d •
🏆 WEEKLY COMP #7: THE OPERATOR 🏆
🎟️ PRIZE: FREE SEAT IN THE LYCEUM 🎟️ Pick your cohort. Technical, Business, or Creator. Your call. ---- 🇬🇧 We're back. Good morning from London. 👋 Thanks for the patience last week. Jake and I needed a few days to breathe before London Tech Week kicked off, and you all responded with nothing but support. We don't take that for granted. Now let's get back to building. ---- 📋 THE CHALLENGE Build a folder-based AI operator that handles ONE operational workflow end-to-end. You pick the workflow. This week's deliverable is one operator folder that someone could drop into a Claude project and use to handle a real business workflow without babysitting. ---- 🎯 PICK YOUR WORKFLOW The workflow is yours. Pick something specific. Pick something you'd actually use. A few sparks to get you thinking: - 🎫 Customer support triage (which tier handles this ticket?) - ✅ Content review and approval - 📨 Lead intake and qualification - 💸 Refund request handler - 🤝 Partnership pitch evaluator - 🎙️ Podcast guest pitch sorter - 💼 Freelance project intake - 📄 Resume screen for one specific role - 📅 Meeting request triage (book, decline, delegate) The more specific, the better. "Customer support" is too broad. "Refund request triage for an ecommerce store doing under 200 orders per month" is right. 📎 If you want a fully written client brief as a reference, the attached PDF walks through one example. Don't build the example. Use it as a template for how to think about scoping your own operator. ---- 🗂️ THE METHODOLOGY If this is your first comp, welcome. Here's what you need to know: This week (and every week) you're learning interpretable context methodology. Folders as architecture. Each file does one job well. Your operator is a folder with five things: - 📄 identity.md (who the operator is and what workflow they own) - 📐 rules.md (the decision logic: criteria, edge cases, escalation rules) - 💬 examples.md (decisions in action, including at least one edge case) - 📚 reference/ (checklists, templates, rubrics) - 📖 README.md (how to use it)
6 likes • 2d
RIPPER (Risk Identification Pattern Prediction Engine Renderer) is an AI-powered risk specialist. It reads structured project context and automatically identifies, scores, and prioritizes risks across six categories using a probability × impact scoring model (1–5 each, yielding scores of 1–25). Every identified risk is assigned to a priority tier — Critical, High, Medium, or Low — and the output is a fully self-contained HTML risk register containing a summary table and per-risk detail cards, each with probability/impact rationale, a mitigation recommendation, a contingency plan, and an owner field. The register is sorted critical-first and grounded strictly in facts from the project brief. Beyond the initial analysis, RIPPER supports interactive register management through conversation: users can challenge scores, add new risks, update statuses, and ask analytical questions like "which risks are most likely to cascade?" It is deliberately narrow in scope — it handles risk analysis only, and when asked about schedules, budgets, scope changes, or stakeholder communications, it names the appropriate future module rather than approximating an answer with risk framing. The project ships with a pre-loaded demo project (the NVX Battery Management System development at KreegCo — a high-stakes BMS replacement with a 9-month hard deadline, OEM contract penalties, IEC 62133-2 certification requirements, and a key engineer departure at Month 5), making it immediately usable. it was tested with 5 different demo projects that where created with different constraints to ensure the logic was sound and edge cases were covered. It is Module 1 of a planned multi-module PM agent. It handles risk analysis only. Github Repo: https://github.com/Kreegs/RIPPER Live Page: https://kreegs.github.io/RIPPER/
3d •
API Tokens
At the rate I am burning through tokens and $$$, I am going to need to find someone to pay me to build something or I am going to need to sell my plasma to keep things going. 🤣 Edit: Issue was poorly written automated test script that was sending the entire folder's contents to the API for each call instead of what was just needed for the pass/fail check.
0 likes • 2d
@David Vogel Here is the worst part, it wasn't even doing iterative checks and improving the logic, it was just doing a basic pass/fair check on the folder logic. I've got notes for the next generation of automated testing to put some guardrails in about api calls what it is doing. I need to be clear on what the testing script needs to do and not do. ICM folders run great and there is hardly any token usage, but between 3 different projects I burned like $50 on API calls for automated testing in the last week. That will teach me: - To reuse code I am not 100% sure about - To double check the code. I do know some python and once I looked I saw what the problem was right away - Not assume the AI model or I didn't hallucinate at some point along the way - Make sure guardrails are in place for the automate testing script development as well as the folders.
0 likes • 2d
@David Vogel I'll write something when I get back from going to my DiL's graduation.
3d •
Do people want AI or Automations but just don't know the difference?
Hi! I am finding a lot of webinars I go to use 'agents', but what it really is, is automations. Anyone else finding the same thing in their field? I work in education and when teachers talk - mostly I want to scream like that is just a workflow - that is just a scheduled task - that is not AI... and then "AI companies" come and want to sell you 'their product' and the same product is a built-in tool?
Poll
8 members have voted
3 likes • 3d
@Curtis Hays This really sums a lot of it up. I had a similar conversation with my boss this week along the same lines. He doesn't understand the difference, but he came to same conclusion out of frustration. He knows what he wants, but nothing out there fits.
2 likes • 3d
@Curtis Hays It would... I wouldn't have spent that last 4 weeks taking all the knowledge around our products from quite literally 23 different sources and compiled into 1 directory containing 2 informational files for every product we make (One technical information, the other marketing information that also points to an assets directory with pictures & videos.) Now if I could get everyone using the same tools, I'd be ecstatic.
4d •
Build Out Question
Hey all looking for architecture advice from anyone who's been here. I'm automating the back-office side of a sales role on Claude — AR/collections, order entry, statement sends, reorder follow-ups, and more workflows being added. The setup is skill-based: I invoke workflows by command (e.g. "Submit orders"), Claude handles the steps, stops for my review before anything sends or submits. Building toward full orchestration with dispatch routing the right workflow automatically based on triggers. Problem: I'm on Max 5x and a few planning chats plus ONE order through Cowork already hit ~26% of my usage before reset. That math doesn't scale once everything is live. The bigger issue: Cowork is currently slower than submitting orders manually. It runs a screenshot-and-click loop through a web UI, and the latency is real. Has anyone found a way to speed this up or get off the vision loop entirely? Questions: 1. Is moving automated execution to the Agent SDK / metered API the right architecture — keep the subscription for interactive work, run agents on the credit pool? 2. For UI-based tools with no API access, did you find a way off the screenshot loop or just accept the cost and speed hit until a native integration exists? 3. Self-capping usage meters on subagents — is that SDK-only or doable another way? 4. Should I be building workflows in Cowork first then graduating to Claude Code, or is it worth starting in Code now? Trying to get the architecture right before I scale this up. Appreciate any wisdom thanks fam.
1 like • 4d
I think your issue is screenshot loop. That is probably chewing up a lot of token use, especially if you are running Fable or Opus. The other stuff really revolves on what programs you are going into and do they have MCP or API's you can leverage? Depending on what you use, there might be APIs that are available that you can then use python to parse the screenshoot then interact with the API with another python script and then hardly any tokens are being used as it all being done locally.
1 like • 4d
@Jordan Tate Yeah, you would build some scripts to run locally to do the work and the ai just provides context. You should see token usage drop like a rock once its built! Good luck and keep us updated!
4d •
New Additions to the Council
Yes, another post about this. Its the topic of the week. I blame @Curtis Hays and @David Vogel for this. Last night I started copying and pasting all of their and @Ari Evergreen's posts and putting them into Obsidian for easy reference. Since Skool doesn't have a save post function, I am just manually doing it now. I am just going to say at this rate each one of them is going with their amazing content, they will probably end up being their own advisors in the council... 😂
1-10 of 112
