Set Up Your Workspace With AI
Use your AI tool to create playbooks and context docs. Just describe what you need in plain English and your agent will build your workspace for you.
Once you've connected Patina to your AI tool, you don't need to use the web UI to set up your workspace. Your agent can create playbooks, add context docs, and link them together — all from a conversation. This guide gives you ready-to-use prompts for the most common setup tasks.
1. Create a playbook
A playbook captures a workflow your team runs repeatedly. If you haven't created one yet, see Create Your First Playbook for an overview. To create one directly from your AI tool, use a prompt like this:
I write [campaign briefs / competitor analyses / call summaries] regularly.
Create a playbook in Patina that captures this workflow. It should include:
- Inputs: [product name, target audience, key messages]
- A structured template that produces a consistent brief every time
- Tips for getting the best results What happens: Your agent calls artifact_put for a playbooks/ path, creating a draft version ready for you to review and approve in the Patina
web UI.
2. Add your brand voice
Context docs let you attach shared knowledge to playbooks — so every run automatically has access to your brand voice, tone guidelines, or internal style. A brand voice doc is one of the most useful things to add early.
Create a brand voice context doc in Patina for our company.
We are [company name], we sell [product/service] to [audience].
Our tone is [professional but warm / technical and direct / casual and friendly].
Include examples of how we write vs. how we don't write. What happens: Your agent calls artifact_put for docs/ with a type: brand frontmatter key, creating a reusable doc you can attach to any playbook.
3. Add a process doc
Process docs describe how your team handles recurring workflows — who does what, in what order, and to what standard. Attaching them to the right playbooks keeps your agents aligned with how your team actually works.
Create a process context doc in Patina that describes how we handle
[content reviews / customer onboarding / sprint planning].
Include the steps, who's responsible at each stage, and any standards we follow. What happens: Your agent calls artifact_put for docs/ with a type: process frontmatter key.
4. What's next: link context to playbooks
Once you have playbooks and context docs, you can link them so every playbook run automatically pulls in the right context. Linking is done in the Patina web UI — open a playbook, go to the Context tab, and attach the docs you want. To get a picture of what you have before you start linking, ask your agent:
List all my Patina playbooks and context docs. Summarize what I have so far
and suggest which context docs should be linked to which playbooks. What happens: Your agent calls artifact_index and artifact_manifest, then gives you recommendations for which docs to link to which
playbooks.
5. MCP tool reference
Power user reference: call tools/list after authentication for the exhaustive, principal-scoped
list. The public server card marks the tool surface as dynamic.
| Read tool | What it does |
|---|---|
artifact_index | Ultra-compact workspace overview (~200 tokens) — every file as path + title |
artifact_manifest | List a directory with metadata (descriptions, links, word counts) but no bodies |
artifact_head | Peek at one file — frontmatter only, or frontmatter plus the first N body lines |
artifact_read | Read a file (markdown with YAML frontmatter); pass raw for round-trippable source |
artifact_search | Full-text search across all visible playbooks and context nodes |
artifact_bundle | Assemble a skill with all its linked business context — call before starting a task |
artifact_links | Follow [[wikilinks]] from a file — what it links to and what links back |
recent | List files changed since a given date |
artifact_history | Show version history for a file |
| Write tool | What it does |
|---|---|
artifact_put / artifact_revise | Create or revise an artifact through the governed contract |
artifact_move | Move or rename a file |
artifact_archive | Archive an artifact through the governed contract |
annotate | Attach a non-canonical agent note to a file |
feedback | Record a quality signal on a file (missing-context, outdated, incorrect, helpful…) |
session_open | Open a session — the agent's workspace for proposing a batch of changes |
session_submit | Submit a session for governed review |
session_cancel / session_revert | Cancel an open session or create a governed revert |