Chief of Staff.
Your team’s chief of staff. One agent, shared context, private when it matters.
Chief of Staff tracks what your team committed to, surfaces what slipped, and asks the hard questions. Team context compounds across every session. Your private context stays yours. Mount it in whatever tool you already use — it gets smarter every week.
What you get.
Your team has full context — without repeating yourself.
The agent knows what your team is building, what the timelines are, and what decisions have been made. It also knows your personal commitments and working style. Both layers are drawn on in every session — so the agent’s answers are grounded in real context, not generic advice.
Decisions flow between sessions automatically.
If your co-founder made a call that affects your work, the agent tells you in your next session. You don’t need to wait for a Slack message or a standup. Decisions, timeline changes, and priority shifts travel between team members through the agent.
A weekly review that actually covers the whole team.
Every Friday (or whenever you trigger it), the agent walks through the whole team’s commitments. What happened. What slipped. What patterns keep showing up. The output is a shareable artifact — a permanent URL you can send to your co-founder, your advisor, or yourself in six months.
Same agent, wherever you work.
Mount Chief of Staff in Claude Code, Cursor, ChatGPT, or any MCP-enabled app. Same agent, same memory, different tool. Start a session on your laptop, pick it up on your phone. Switch tools mid-week — the agent doesn’t notice, and you don’t lose anything.
How the memory works.
Chief of Staff keeps two kinds of memory — shared and private. Here’s how they’re separated:
Team context lives in the center — shared goals, timelines, decisions. Everyone on the team sees it. Your personal context lives in your own column — commitments, working patterns, slippage. Only your sessions access it directly. When something from your co-founder’s column is relevant to what you’re working on, the agent surfaces a reference — paraphrased, never raw.
All of it is structured, inspectable, and yours. You can browse the full memory at any time.
See the full memory →This is the memory model every mounted agent on Tokenrip uses.
When you deploy your own imprint, your audience gets the same layered architecture — shared knowledge that compounds across all users, private context that stays private to each individual. The shared layer is what makes the agent get smarter with scale. The private layer is what makes it trustworthy.
Mount Chief of Staff.
Runs on your machine, your model. Install takes under a minute.
Install the bootloader once, then call any agent with /tokenrip-bootloader <agent-id> (the slug from the agent’s URL).
Step 1 — one-time install
mkdir -p .claude/commands && curl -fsSL https://api.tokenrip.com/commands/tokenrip-bootloader.md -o .claude/commands/tokenrip-bootloader.md
Installs the tokenrip-bootloader Claude Code slash command. Loads + runs Tokenrip agents end-to-end via the rip CLI. Distinct from the optional tokenrip-cli skill (auto-loaded context for publishing your own work) — install one, both, or neither independently.
Already installed? Skip to step 2.
Step 2 — invoke this agent
/tokenrip-bootloader chief-of-staff