II.
Page overview
Reference · livepage:agent-generate-universal-agentic-stack-placement-checklist
Placement Checklist overview
Inspect the raw attributes, linked wiki pages, and inbound or outbound graph edges for page:agent-generate-universal-agentic-stack-placement-checklist.
Attributes
nodeKind
Page
title
Placement Checklist
displayName
Placement Checklist
slug
agent-generate/universal-agentic-stack/03-placement-checklist
articlePath
wiki/agent-generate/universal-agentic-stack/03-placement-checklist.md
article
# Placement Checklist
Use this checklist when someone says "this tool is an agent platform" or "this framework does everything." It forces the classification into concrete responsibilities instead of labels.
## Step 1: Pin down the model side
- What model family or version is actually being used?
- Who serves it?
- What transport or protocol carries requests to it?
This covers Layers 1, 2, and 3.
## Step 2: Pin down the agent side
- Where is the actual decision loop or graph logic?
- Where do tools, approvals, interrupts, and session state live?
- Where do installed skills, plugins, channels, and extension contracts live?
This covers Layers 4, 5, and 6.
## Step 3: Pin down the operating boundary
- What workspace does the agent read and write?
- Where do shell commands, browser runs, or tool actions execute?
- What policy restricts files, network, secrets, or binaries?
This covers Layers 7, 8, and 9.
## Step 4: Pin down the surface
- What actions are exposed to users or systems?
- How are results rendered, streamed, or integrated?
This covers Layers 10 and 11.
## Step 5: Record what is delegated
Not every product owns every layer. Explicitly note when a layer is delegated to:
- an IDE
- a host application
- a CI runner
- a cloud control plane
- a user-managed shell or container
Delegated layers still exist. They just are not owned by the product you are classifying.
## Step 6: Write the placement in one sentence
Force yourself to write a short sentence with three parts:
1. main owned layers
2. main delegated layers
3. main exposed surface
If you cannot do that clearly, your classification is still too fuzzy.
## Minimal worksheet
| Layer band | Owned | Exposed | Delegated | Notes |
|---|---|---|---|---|
| 1-3 Model supply | | | | |
| 4-6 Agent system | | | | |
| 7-9 Operating boundary | | | | |
| 10-11 Surface | | | | |
Fill this before arguing about labels.
## Output format
When you finish, you should be able to say something like:
`This product mainly owns Layers 4-6, delegates 7-9 to the host environment, and exposes Layers 10-11 through a web UI and API.`
## Follow-up
- Use [`layers.md`](./layers.md) to verify disputed boundaries.
- Use [`04-worked-examples.md`](./04-worked-examples.md) if you want to compare your result to common product shapes.
documents
Outgoing edges
documents11
- layer:1-model·LayerModel
- layer:2-provider·LayerProvider
- layer:3-transport·LayerTransport
- layer:4-agent-core·LayerAgent-Core
- layer:5-agent-runtime·LayerAgent-Runtime
- layer:6-agent-platform·LayerAgent-Platform
- layer:7-workspace·LayerWorkspace
- layer:8-execution·LayerExecution
- layer:9-sandbox·LayerSandbox
- layer:10-interaction·LayerInteraction
- layer:11-presentation·LayerPresentation
Incoming edges
contains_page1
- page:agent-generate-universal-agentic-stack·PageUniversal Agentic Stack