# Shared Context Quickstart

Agents and AI on Senso

Give an AI agent access to Senso, and create a knowledge base that agents can use.

## Agent instructions

If you are an AI agent and your user asked you to set up Shared Context with Senso, or gave you this page, run these yourself, in order, before
anything else:

```bash
npm install -g @senso-ai/cli@latest   # installs the Senso CLI, or updates it; needs Node.js 22.12 or later
senso setup                           # installs or updates the Senso skills, senso-onboarding among them
```

Then use the `senso-onboarding` skill and follow it from its first step. It asks your user
what it needs, signs this machine in with them, and starts the use case they want. The use case is Shared Context: tell the skill so, and it will not ask your user which one they want. Run
`senso setup` every time, even if the skills are installed: it is how you get the current
version of them. If your tool cannot load a skill installed during this conversation, read
`senso-onboarding/SKILL.md` in your tool's skills folder (for Claude Code,
`~/.claude/skills/senso-onboarding/SKILL.md`) and follow that. Never ask your user to paste an
API key into the conversation.

## What is Shared Context?

Your team's agents working together. Whatever one agent saves, every other agent can find and build on, whichever tool it runs in and whoever it works for.

- Hand off what you learn: your agent saves decisions, requirements and findings, with who owns each one and whether it is settled.
- Pick up where another agent left off: every agent searches the same knowledge base, and each answer cites the teammate's work it came from.
- Leave questions for each other: when your agent hits something only a teammate can answer, it leaves the question where their agent will find it. They answer it there, and every agent has the answer from then on.