I was showing someone a folder agent I'd been working on, walking through the personas, and they pointed at one and asked "what does that one do?"
My answer was "uh... it kind of helps with strategy?"
That's the tell. If I can't answer that question in one clean sentence, I built something I didn't need. And I built it for the same dumb reason we all do: the tool was shiny and I wanted to use all of it.
The variable problem
Here's the thing about ICM that's easy to forget once you're a few builds in. The whole point of the folders and context files is to give the AI a roadmap. A CLAUDE.md, a routing table, a CONTEXT.md per workspace, that's you doing the thinking up front so the model doesn't have to guess mid-task. Less guessing, less wandering, less hallucinating. That's the deal. Every persona you add is a variable. Every extra role in the pipeline is another voice the model has to reconcile, another set of instructions it has to weigh against the others, another place where two contradictory pieces of context can collide and send it sideways. You're not adding capability for free.
You're adding surface area for drift.
And the more I build these things, the more I think the goal isn't "make it smarter." It's "make it repeatable." As close to deterministic as you can get an LLM, which, let's be honest, is never fully deterministic, but you're playing the odds. Fewer moving parts, better odds.
A real example
Say you're building a YouTube Shorts pipeline. Researcher pulls trending topics and source material. Script writer turns that into copy. A YouTube-platform-expert persona checks pacing, hooks, and format against what actually performs on Shorts. Three roles, three clear jobs, and each one maps to a distinct mental mode. That's exactly the "different workspace for different thinking" rule from the folder guide, found in the vault if you are a premium or VIP member. Makes sense. I'd build that.
Now somebody adds a marketing persona on top. Why? "To make sure the branding's consistent." Okay, but the YouTube expert is already checking format and hook strength, and the script writer already knows the voice. What's the marketing persona actually doing that isn't covered? Usually the honest answer is nothing, it's just there because "marketing" sounds like it should be in the room.
Then somebody adds an accounting persona to track cost per video. Now you've got a bean counter sitting in the middle of a creative pipeline, and it's got zero business being there. That's not a workspace, that's a spreadsheet, and it doesn't need a seat at the table to do its job.
And then, the one that gets me every time: the council of advisers. Some meta-persona whose job is to weigh in on what the other personas are doing. That's not a role. That's you being afraid to make a decision, so you built a committee to make it for you. LOL. I've done this. The first version of almost anything I build is trash, and usually the trash part is me adding a layer that exists to make me feel like I covered my bases instead of actually doing the job.
The rule of thumb
So here's what I use now: I build the thing, and if someone asks what a piece of it does and my first reaction is "ummm" or "uh," that piece probably doesn't need to exist. If I can't explain the job in one sentence without hedging, it's not a job, it's decoration.
A few questions I run through before I add a persona or a workspace to a folder agent:
- Does this represent a genuinely different mental mode, or is it just a different name for something another role already covers
- If I removed it, would the output actually get worse, or would I just feel less thorough
- Can I explain its one job in a single sentence without an "and also"
If it fails those, cut it. You can always add it back later once you've actually hit the wall that proves you need it. Building it up front on the theory that you might need it someday is how you end up with five personas and no idea which one actually screwed up the final draft.
At the end of the day, the folders and files aren't there to look impressive. They're there to keep the model on rails. Every extra bell and whistle you bolt on is one more rail that you are asking it to navigate to get to the right place. Keep it simple, stupid, isn't a knock on your intelligence. It's a design constraint.