streamlinear MCP
A Linear MCP server with one tool and seven actions that takes about 500 tokens of context against 17,000
- 90 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The Linear API token opens reading and writing issues, comments and arbitrary GraphQL queries in your workspace
- The `graphql` action runs any mutation within the token permissions
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/streamlinearDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The Linear API token opens reading and writing issues, comments and arbitrary GraphQL queries in your workspace
- The `graphql` action runs any mutation within the token permissions
This is third-party code. Review the repository files before installing.
What it does
In place of 23 separate tools streamlinear gives the agent one tool with an `action` parameter: `search`, `get`, `update`, `comment`, `create`, `graphql` and `help`. According to the author, the standard Linear MCP server definitions take about 17,000 tokens and streamlinear about 500, which leaves context for the work. Defaults keep requests short. A search with no parameters returns your active issues without completed and canceled ones. An ID is accepted as ABC-123, a Linear URL or a UUID. State names match fuzzily (in prog gives In Progress) and `assignee: "me"` points to you. Teams and workflow states load at startup and appear in the tool description. The `STREAMLINEAR_POST_ONLY=1` mode forbids reads (`search`, `get`, `graphql`) and keeps create, update and comment for cases where the agent must not see issue content.
Who it is for. Teams on Linear who connect an agent to the tracker and count context tokens.
Good fit when
- The agent must find your active Linear issues and move one to Done
- You need to create an issue and add a comment from the session with no extra tool descriptions in context
- You need GraphQL queries to Linear for things the ready actions do not cover
- Confidentiality requires the agent to write to Linear and never read issue content
Not a fit when
- You want separate tools with detailed schemas for each operation: this has one shared tool
- The work runs in Jira, YouTrack or another tracker
- You have no Linear API token or do not want to give one to an agent
Example request
Show my active Linear issues and move ENG-142 to Done with the comment fixedMCP
- Transport
- stdio
- Authentication
- API key
Environment variables
LINEAR_API_TOKENrequired, secret- Personal Linear API token for workspace access
STREAMLINEAR_POST_ONLY- Value 1 forbids reading issues and keeps writing
Limitations
- A personal Linear API token in the `LINEAR_API_TOKEN` variable is required and gives access to your workspace.
- All capabilities sit in one tool with an `action` parameter, so the schema is less detailed than a set of separate tools.
- Multiple workspaces need separate MCP entries with separate tokens.
- The `graphql` action accepts arbitrary queries and mutations, and their safety depends on the token permissions.
How to disable. Remove the `linear` entry from `.mcp.json` or run `/plugin uninstall streamlinear` in Claude Code. Revoke the token in Linear settings.
FAQ
How do I connect streamlinear?
Add a server to `.mcp.json` with the `npx` command, the arguments `-y` and `@primeradianthq/streamlinear@1.2.1` and the `LINEAR_API_TOKEN` variable. A Claude Code plugin variant with the same MCP also exists.
How many tokens does the single tool save?
According to the author, about 500 tokens against about 17,000 for the standard Linear MCP server with 23 tools.
How do I use several Linear workspaces?
Create separate MCP entries such as `linear-personal` and `linear-work` and map each secret to `LINEAR_API_TOKEN`.
Can I stop the agent from reading issues?
Yes, `STREAMLINEAR_POST_ONLY=1` disables `search`, `get` and `graphql`. Create, update, comment and help keep working.
What does the `graphql` action do?
Runs an arbitrary GraphQL query against Linear when the ready actions are not enough. `{"action": "graphql", "graphql": "query { viewer { name } }"}` returns the token owner's name.
Discussion
No questions or reviews yet.
Related
MCP serversAn MCP server for Linear: issues, projects, cycles, milestones, documents and project updates via the GraphQL API
MCP server for Linear: the agent creates and updates issues, searches the tracker and adds comments through the Linear API
Linearis manages Linear issues and projects from the terminal, returns JSON and exposes focused command references to agents.
Linear CLI
linear-cli
A Linear CLI for the terminal: list, start and create issues, PRs from a branch, JSON output and an agent skill, with git and jj support
Go CLI for the Linear API built with agents in mind: issues, projects, comments, raw GraphQL, JSON output and a built-in MCP
A self-hosted issue tracker and wiki on Cloudflare with MCP: 126 tools, issue leases and file claims for several agents
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| streamlinear (this card) | no | API key | 90 | |
| Linear MCP Server | no | API key | 346 | separate tools with a detailed schema per operation, while this server folds everything into one tool with an action parameter |
| MCP Linear | no | API key | 151 | covers issues, projects, cycles and documents with a tool set, while this server saves context with one tool |
README in short
The README explains the idea of a single tool with action dispatch and gives a table of seven actions. Installation for one and several workspaces, the package binaries, post-only mode and the defaults follow: fuzzy state matching, parameterless search and different ID forms.