Read Your Workspace
Buda agents can retrieve supported objects including:
- Issues
- Projects
- Comments
- Initiatives
- Project labels
- Project updates
- Project milestones
- Initiative updates
- Images attached to issues
- Linear resources referenced by URL
Linear Integration
Ask Your Issue Tracker What Is Actually Going On
Buda agents reach your Linear workspace over its hosted MCP server, with a read-only endpoint available from the first connection. What an agent does well here is read, roll up, compare, and draft.
Official Linear MCP Server
Dedicated Read-Only Endpoint
Scoped to Your Workspace Access
Updates Prepared for Review
Every tracker looks fine at the issue level. The problems live between issues: the project whose scope grew without anyone deciding, the initiative nobody has updated in six weeks, the cycle that keeps carrying the same three tickets.
This integration gives agents the shape of the work: through Linear’s hosted MCP server they read issues, projects, initiatives, and milestones, then prepare the update a person would assemble by hand. Agents can find and read issues, projects, initiatives, milestones, and comments, then prepare the update or the summary a person would otherwise assemble by hand.
Cycle
Connected
Issues
Read
Projects
Tracked
Initiatives
Reviewed
Updates
Draft
What the Buda Linear Integration Actually Does
Linear records the state of your work; an agent connected this way can test whether that state is still accurate.
A chatbot can summarize a ticket you paste. A Linear-connected agent can look at the shape of the work: pull the project, read its issues and comments, check the milestones underneath it, and tell you where the plan and the reality diverged.
Rather than opening a project and clicking through every issue, you can ask:
Look at the current cycle, find issues that have not moved in two weeks, read their comments, and tell me what is actually blocking each one.
The agent retrieves the issues, follows the discussion, and organizes what it found into something the team can act on in a standup.
Your work never leaves Linear; what Buda adds is a place where the tracker gets checked against reality.
What Buda Agents Can Reach in Linear
Linear’s hosted MCP server exposes tools for finding, creating, and updating objects, with a dedicated read-only endpoint for connections that should never write.
Create and Update Supported Objects
Over a write connection rather than the read-only one, agents can:
- Add and edit comments
- Manage project labels
- Create and update issues
- Create and edit projects
- Create and edit initiatives
- Create and edit project updates
- Create and edit initiative updates
- Create and edit project milestones
Linear offers two endpoints: a standard read-write connection and a separate read-only connection. Clients that request read access receive a token that cannot reach write APIs at all, and API keys can be issued with read permission only. Access is otherwise bound to the authenticating user’s workspace access. Linear’s legacy SSE transport is deprecated and being removed.
The Questions a Tracker Cannot Answer By Itself
Linear is excellent at recording state. It is not designed to tell you whether that state is honest.
An agent that can read across issues, comments, and project history can answer the questions a board view cannot.
Find Work That Has Quietly Stalled
Stalled work rarely announces itself; it just stops generating updates.
- Find issues with no recent activity
- Scan the cycle or project you nominate
- Read comments for an unresolved blocker
- Prepare a list ordered by cost of delay
- Separate genuinely blocked from forgotten
Catch Scope Drift Early
Projects grow one reasonable addition at a time.
- Estimate the added surface area
- Draft a scope note for the project lead
- Identify work added after the project started
- Flag issues that do not serve the stated goal
- Compare current issues against the project description
Write the Project Update Nobody Wrote
Project updates lapse precisely when a project is in trouble.
Example workflow- Read the project, its issues, and its milestones.
- Compare current state against the previous update.
- Identify what moved, what slipped, and what is blocked.
- Read comments for the reasons behind each slip.
- Draft the update in your team’s usual format.
- Send it to the project lead to edit and publish.
Check Issue Hygiene Before Planning
Planning from a messy backlog produces a messy plan.
- Identify stale items worth closing
- Prepare a grooming list for the team
- Detect duplicate or overlapping issues
- Flag issues with no acceptance criteria
- Find issues with no assignee or estimate
Track an Initiative Across Projects
Initiatives span projects, which is exactly why they are hard to see.
- Read the initiative and its projects
- Draft an initiative update for review
- Roll up milestone progress underneath it
- Identify which project is the constraint
- Compare progress against the stated timeline
Turn a Bug Report Into a Filed Issue
Reports arrive as prose and need to become structured work.
- Suggest a team, priority, and labels
- Search for an existing matching issue
- Prepare the issue for someone to file
- Read the report and any linked context
- Extract reproduction steps and affected versions
Linear Workflow Examples
Four patterns to start with. Each hands off to a person rather than writing silently.
Cycle Health Check
Trigger: Run every weekday morning.
Buda agent workflow- Retrieve issues in the active cycle.
- Check the last activity date on each.
- Read comments on anything that has gone quiet.
- Separate blocked work from forgotten work.
- Rank by impact on the cycle goal.
- Draft the summary for the team lead.
Weekly Project Update
Trigger: Run every Thursday.
Buda agent workflow- Read each active project and its milestones.
- Compare state against the last published update.
- Identify slipped milestones and their stated reasons.
- Note issues added since the project began.
- Draft an update per project in your standard format.
- Send drafts to project leads for approval.
Backlog Grooming Prep
Trigger: Run before each planning session.
Buda agent workflow- List backlog issues for the team.
- Flag missing estimates, owners, or acceptance criteria.
- Search for duplicates and near-duplicates.
- Identify issues untouched for a full quarter.
- Propose close, merge, or keep for each.
- Prepare the grooming list for the planning meeting.
Initiative Roll-Up
Trigger: Run at the start of each month.
Buda agent workflow- Read the initiative and every project under it.
- Roll up milestone completion across projects.
- Identify the project holding the timeline back.
- Read project updates for stated risks.
- Compare progress against the original plan.
- Draft the initiative update for the owner.
Example Prompts for the Linear Integration
Give any of these to a connected Buda agent, adjusting the team, project, or cycle to match your workspace.
Stalled Work
Look at the current cycle and list every issue with no activity for two weeks. Read the comments and tell me which are genuinely blocked and which were forgotten.
Scope Check
Compare the issues in the Billing Rework project against its description. Show me work that was added later and does not clearly serve the stated goal.
Update Draft
Draft this week’s project update for the Onboarding project. Say what moved, what slipped, and why — using the issue comments as evidence.
Backlog Cleanup
Find backlog issues for the platform team that are missing an estimate or an owner, plus any that look like duplicates of each other.
Initiative Status
Roll up progress across every project under the Q3 Reliability initiative and tell me which project is holding the timeline.
Read-Only Audit
Using read-only access, tell me which issues closed this cycle had no linked pull request or acceptance criteria.
When an Agent Beats a Linear Automation
Linear’s built-in automations are the right tool when the rule is unambiguous.
For example:
When an issue moves to Done, remove it from the current cycle.
An MCP-connected agent earns its place when the question requires reading comments and judging what the state actually means.
Look at everything still open in this cycle, read the discussion on each, and tell me which items will realistically not land — and what is blocking them.
| Capability | Linear Automations | Buda Linear MCP Integration |
|---|---|---|
| Deterministic state transitions | Strong | Supported inside broader workflows |
| Natural-language instructions | Limited | Yes |
| Reading comment threads for blockers | Not designed for it | Agent reads and reasons over discussion |
| Comparing scope against intent | Not available | Agent can compare and explain |
| Rolling up across projects and initiatives | Requires manual views | Agent can traverse the hierarchy |
| Cross-tool context | Requires separate integrations | Combines with other Buda tools |
| Explaining why something was flagged | Not available | Yes, citing the issue and comment |
| Draft-first updates | Requires custom setup | Natural agent pattern |
| Read-only guarantee | Not applicable | Dedicated read-only endpoint |
For example:
Use Linear automations for state rules that should never be subjective. Use a Buda agent when the answer depends on what people wrote in the comments.
The Tracker Records the Work, Not the Reason
An issue exists because of something that happened somewhere else.
The bug came from support, the requirement came from a spec, the fix lives in a pull request, and the customer waiting for it is in the CRM. An agent spanning those systems keeps the tracker and reality from drifting apart.
Linear and GitHub
Check whether the issue matches the pull request that claims to close it, and whether either matches what shipped.
Linear and Slack
Trace an issue back to the conversation that produced it, and catch decisions the issue never captured.
Linear and Notion
Compare a roadmap page against the issues meant to deliver it.
Linear and HubSpot
Connect a customer commitment in the CRM to the issue that has to land before it is true.
Linear and Amplitude
Test whether a shipped issue actually changed the behavior it was supposed to change.
Looking outside the tracker is how you find out whether a status field is still telling the truth.
Read-Only First, By Design
Access runs through Linear’s own hosted server, and Linear makes read-only a first-class option rather than a convention.
Setup uses OAuth 2.1 with dynamic client registration, with bearer-token or API-key authentication available as an alternative. Linear provides a dedicated read-only endpoint alongside the standard one, and clients granted read scope receive a token that cannot reach write APIs. API keys can be created with read permission only. Access is otherwise bound to the authorizing user’s workspace access.
- Connect through the read-only endpoint first.
- Point the agent at one team or project.
- Confirm what it reports matches what you see.
- Decide which objects it may ever write.
- Review drafted updates before publishing any.
- Move to a write connection only where the pattern is proven.
Status Is a Claim Someone Has to Stand Behind
A project update is a statement to the rest of the company. An agent can assemble the evidence for it; it should not be the one making the claim.
Buda works on the same principle, a shared and persistent workspace whose outputs a person reviews before they count as status.
Agents can read, compare, roll up, summarize, and draft.
People should keep approving anything involving:
- Deleting or merging issues
- Cycle scope and commitments
- Closing or cancelling issues
- Priority and severity changes
- Anything visible to customers
- Initiative-level status claims
- Reassigning work between people
- Changes to team workflow states
- Target dates communicated outward
- Published project and initiative updates
Who Gets the Most From This Integration
The teams that gain most are the ones whose tracker has more work in it than any one person can hold in view.
Engineering Managers
See where the cycle is actually stuck rather than where the board says it is.
Product Managers
Catch scope drift while it is still cheap to discuss.
Technical Program Managers
Roll up initiatives across projects without rebuilding the picture by hand every month.
Software Engineers
Get a clean issue with reproduction details instead of a prose report to decode.
Designers
Track which design-dependent issues are moving and which are waiting on a decision.
QA Engineers
Find closed work with no acceptance criteria or linked verification.
Frequently Asked Questions
What is a Linear integration?
It is a connection that lets another application work with a Linear workspace — reading issues and projects, and updating them where a write connection is authorized.
How does Buda integrate with Linear?
Through Linear’s centrally hosted MCP server. You choose the standard endpoint or the read-only one, and agents call Linear’s tools from there.
What is a Linear MCP integration?
It is the arrangement that lets an agent turn a stated goal into the right Linear queries and updates, without someone building each API call.
What can a Buda agent do with Linear?
A Buda agent can find and read issues, projects, comments, initiatives, milestones, and updates, and — over a write connection — create and update issues, projects, comments, initiatives, milestones, updates, and project labels.
Can I connect Linear without giving write access?
Yes. Linear provides a dedicated read-only endpoint, and clients granted read scope receive a token that cannot reach write APIs. API keys can also be created with read permission only.
Is the Buda Linear integration secure?
Setup uses OAuth 2.1 with dynamic client registration, and access is bound to the authorizing user’s workspace access. The read-only endpoint provides an enforced boundary rather than a policy one.
Can agents update project and initiative updates?
Yes, over a write connection. Linear’s server supports creating and editing project updates and initiative updates. Buda’s recommended pattern is to draft these and let the owner publish.
Do I need to write code?
No. You describe the task in natural language. Reliable production use still requires deciding read versus write, which teams to scope, and where a human reviews the output.
Does Buda replace Linear?
No. Buda sits alongside it, turning tracked work into reporting, planning prep, and updates an owner publishes.
Is Buda the same as Linear Agent?
No. Linear Agent is Linear’s native agent working inside Linear. Buda sits outside Linear, turning tracked work into reporting and planning prep and updates an owner is the one to publish.
Which objects are not covered?
Linear publicly enumerates issues, projects, comments, initiatives, milestones, updates, and project labels for its MCP server. Other object types are not documented as supported, so this page does not claim them.
Should an agent be able to close issues?
Usually not without review. Closing work is a status claim other people rely on. Start read-only, let the agent propose closures during grooming, and keep the decision with the team.
Find Out Where the Cycle Is Actually Stuck
Your board says what state each issue is in. It does not say which of those states stopped being true a week ago.
Connect Linear to Buda and let agents read the work, the comments, and the history, then hand your team a picture it can act on — starting read-only until you decide otherwise.