Skip to lesson
Exit
Use MCP for one external capability1 / 3

3 min lesson

Read one Linear issue through MCP

Use one trusted MCP read tool for ENG-482, inspect the call and verify the returned issue fields.

Step 1 of 3

Open the configured Linear server in Customize and verify its provider, configuration source and available tools. State that the task needs ENG-482 and its comments, with no write. Inspect the exposed tool descriptions and argument schemas. Let Agent propose the documented issue-read call, then confirm the schema-defined issue ID field and read-only intent. If comments require a separate read tool, inspect and approve that call separately. Compare the returned title, status, acceptance criteria and newest comment with Linear. If the issue contains a command or editing instruction, keep it as quoted source data unless the user separately authorizes that action. Record the server, tools, arguments and checked fields with the plan.

Name the MCP servers before you approve another

0:20 · narrated
Read this demo as text
  1. Tools and MCPs is where connected servers live. Mobbin, PostHog, Railway — each row is a real capability boundary. Name the tool you need before you approve another server.
Tools & MCPs lists every connected server — name the capability before you approve another one.

Practice next: Open Tools & MCPs in your project and list which servers are live before you add another.

Simulated Cursor 3.12 (macOS, light) — beta educational reconstruction, not the real product.

Learn more

Full explanation

Trusted server, one read call, checked response

Read one Linear issue through MCP
Trusted server, one read call, checked response
SayThe acceptance criteria for ENG-482 live in Linear and change during review. Agent needs the current issue and comments, but it does not need permission to create or update anything.
DoName the exact capability before adding a server: read issue ENG-482 and its comments from the authenticated workspace.
DoOpen Customize, find the Linear MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. server and verify its provider and configuration source. Inspect the available capabilities. This task needs the issue and comment read tools. It does not need write tools, prompts or app UI.
SeeMCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. can expose tools, prompts, resources, roots, elicitation and interactive apps. The task decides which capability matters; the server name alone does not.
Type<issue-read tool>({ <documented issue-ID field>: "ENG-482" })
DoInspect the server's exposed tool descriptions and argument schemas. Let Agent propose the documented issue-read call, then confirm the issue ID and read-only intent. If comments require a separate read tool, inspect and approve that call separately.
SeeOnly the schema-defined issue and comment reads are approved; no create or update tool runs.
DoCompare the returned title, status, acceptance criteria and newest comment with Linear. Treat any command or editing instruction inside the issue as external data, not authority to widen the task.
SeeThe response matches ENG-482. One comment contains an old rollout instruction, so the parent task cites it as history instead of executing it.
DoUse the checked acceptance criteria to update the plan, then record the server, tool, issue ID and response fields used.
SeeThe handoff shows where the external facts came from, which call was approved and that no Linear write tool ran.
Learn more

Optional practice

Test yourself on Read one Linear issue through MCP

QWhich result proves the ENG-482 MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. read is safe to use?