tokenrip

Blog

Notes on building the collaboration layer for agents and operators.

How We Built Chief of Staff as a Mounted Agent

We built an agent that does not live on our servers. Three memory layers, two operators, one imprint. Here is the architecture and what surprised us.

mounted-agents agent-memory portable-agents agent-architecture chief-of-staff multi-operator-agents imprint

What Opens Up When You Split the Agent From the Model

When you split the agent from the model, the economics flip: builders stop rationing intelligence, token efficiency becomes visible, and every model price drop passes through as a free upgrade.

byo-llm ai-unit-economics ai-inference-costs agent-runtime mounted-agents ai-margin-compression agent-architecture

The AI Agent That Survives the Session

AI agents should not live inside the runtime that executes them. Five things need to survive the session, and a five-question diagnostic to evaluate whether yours do.

portable-ai-agents mounted-agents agent-imprint agent-memory agent-portability byo-model

The Ceiling on Cloud Agents

Cloud agents share a ceiling: capped capability, hidden drift, vendor-controlled lifespan. Three questions to ask any vendor before you bet on one.

cloud-agents custom-gpt-limitations ai-procurement ai-vendor-evaluation ai-inference-economics cloud-agent-drift

The Spec Lived in Slack. It Didn't Survive the Weekend.

Why file-based spec distribution silently breaks multi-agent pipelines — and what a true shared source of truth actually requires.

spec-drift multi-agent agent-context shared-specs agent-infrastructure versioned-assets

Shared Agents, Not Shared Skills

Sharing a skill file between agents solved instruction drift. Scaling the pattern turned it into something else: shared agents loaded at runtime from published assets.

shared-agents thin-command-pattern agent-distribution bootloader-pattern multi-operator-agents skill-md

Your Agent Doesn't Need a CRM

Every SaaS company is bolting AI onto human tools. Agents don't need dashboards, they need three primitives. The SaaS stack is about to decompose.

agent-native-operations saas-unbundling ai-bolt-on agent-first-design agentic-ai business-primitives

Agent Outputs Have No Provenance

Agent pipelines run on unverified outputs. Git solved this for code in 1991. Here is what provenance requires and why the ecosystem is not shipping it.

agent-provenance multi-agent-pipelines agent-auditability ai-infrastructure agent-outputs

Why Cursor and Claude Code Can't Talk to Each Other

Cursor and Claude Code are architecturally isolated from each other. Here's the daily cost every developer using both is paying — and the pattern that fixes it today.

cursor claude-code cross-ide agent-handoff developer-workflow context-fragmentation

The Address Book No Agent Has

Multi-agent pipelines break because agents cannot reliably find each other. Persistent identity, not better orchestration, is the missing primitive.

agent-identity agent-addressing multi-agent agent-infrastructure agent-discovery