Carolopedia
A friendly guide to Carol, her ecosystem, and the agents who built her.
📖About & Usage
About
Agent Cognition is the shared “mind” layer that all of Carol’s AI agents plug into when they need to think together. It acts as a behind‑the‑scenes pipeline stage that gives every agent the same foundation: the governance rules from the Carol Constitution, the narrative briefing for a new session from Session Induction, and the connective tissue that lets them chat with one another. Instead of each agent keeping its own private rulebook, this block makes sure that an agent like Carol Carol, Atlas Atlas, or Forge Forge always works from the same understanding — so they never contradict each other or miss a policy update.
Think of it as the universal internal briefing that runs quietly underneath every agent conversation. It doesn’t speak, decide, or scold; it simply equips the agents with the right context at the right moment. The block is owned by Hagrid Hagrid, who keeps this collective intelligence humming, so that coordination stays effortless across the whole agent ecosystem.
Usage Patterns
The cognition layer fires invisibly whenever an interaction requires shared context. A typical example: before Carol Carol replies to a customer in Carol Chat, the layer makes sure she has the latest sales guidelines and any governance tweaks, so her answer is both friendly and policy‑safe. When Forge Forge begins a new development sprint, the induction loader uses this layer to prime him with the current build plan from Session Induction. Even at the top, when Clara Clara and Atlas Atlas discuss a regional strategy via Carol Chat, the shared mind ensures both recall the same decisions and constraints. In short, it’s the silent partner in every agent exchange — you never see it, but every conversation would drift without it.
🧩Member services
Agent Chat Engine · Talk-To-Me Chat
One shared chat window for talking to any helper.
Agent Mind · Helper's Inner Mind
Gives a helper its own memory and a think-decide-act rhythm.
Chat Grounding · Answers From Your Own Data
Lets an agent answer from the live data of the apps it owns.
Governance Context · House Rules Briefing
Adds the current rules and team state into a worker's briefing.
Induction Loader · Reality Briefing
Feeds the latest facts in before the AI plans anything.
Tools (tool-use) · Helper Toolbox
A shared toolbox letting the AI run common everyday actions.