Linear Action Nodes

Linear action nodes let workflows read and update Linear issues, create comments, manage AgentSession state, and perform text manipulation on Linear content. All Linear action nodes use your organization's connected Linear OAuth credentials - no API key configuration is needed in individual node configs.

Linear uses team-canonical state names. The state "Started" is the correct name for active work on most Rensei teams - "In Progress" silently no-ops. Always verify state names in your Linear team settings before configuring linear.issue.update or linear.issue.advance.

AgentSession nodes

`linear.agent_session.acknowledge`

Acknowledges a Linear AgentSession, signaling to Linear that your workflow has accepted the work. Call this as the first action after a linear.agent_session.created trigger to prevent the session from timing out in Linear's UI before your agent begins.

Input ports

Port	Type	Required	Description
`sessionId`	string	yes	Linear AgentSession ID (from trigger output)
`organizationId`	string	no	Override org context

Output ports - success (acknowledged) / fail ({ error: string })

`linear.agent_session.close`

Closes a Linear AgentSession by emitting a response activity that transitions the session state from active to complete. Place this at the terminal branch of every SDLC stage to give users a clean session close in the Linear UI.

The Linear API has no explicit closeAgentSession mutation. Rensei implements close by emitting a response activity, which transitions the AgentSession to complete state. Sessions left without a close call appear perpetually active in the Linear interface.

Input ports

Port	Type	Required	Description
`sessionId`	string	yes	Linear AgentSession ID
`reason`	string	no	Optional closing message shown in the Linear session thread
`organizationId`	string	no	Override org context

Output ports

Port	Field	Description
`success`	`{ closed: boolean, skipped?: boolean, reason?: string }`	Session was closed (or suppressed if already closed)
`fail`	`{ error: string }`	Linear API error

skipped: true is returned when the session was already closed (idempotent). Sessions that were never forwarded to Linear (self-dispatched stages with synthetic UUIDs) are also skipped silently - no error is raised.

Issue nodes

`linear.issue.read`

Fetches the current state of a Linear issue by ID.

Input ports

Port	Type	Required
`issueId`	string	yes

Output ports - success (LinearIssue) / fail ({ error: string })

LinearIssue shape

{
  id: string
  identifier: string    // e.g. "REN-1234"
  title: string
  description?: string
  url: string
  state: { id: string; name: string; type: string }
  assignee?: { id: string; name: string; email: string }
  labels?: Array<{ id: string; name: string; color: string }>
  team: { id: string; name: string; key: string }
  priority?: number
  estimate?: number
}

`linear.issue.update`

Updates one or more fields of a Linear issue.

Input ports

Port	Type	Required	Description
`issueId`	string	yes	Linear issue UUID
`title`	string	no	New title
`description`	string	no	New description (Markdown)
`status`	string	no	New state name (must be a canonical team workflow state)
`assigneeId`	string \| null	no	Assign to user UUID; pass `null` to unassign
`labelIds`	`string[]`	no	Replace label set
`estimate`	number	no	Story point estimate
`priority`	number	no	Priority (0=none, 1=urgent, 2=high, 3=medium, 4=low)

Output ports - success (LinearIssue) / fail ({ error: string })

`linear.issue.advance`

Advances a Linear issue to its next workflow state. Uses the team's workflow state ordering - you specify the issue ID, and the platform resolves the next state automatically.

Input ports

Port	Type	Required
`issueId`	string	yes
`organizationId`	string	no

Output ports - success (LinearIssue) / fail ({ error: string })

`linear.issue.create_sub_issue`

Creates a new sub-issue under a parent Linear issue.

Input ports

Port	Type	Required	Description
`parentIssueId`	string	yes	Parent issue UUID
`title`	string	yes	Sub-issue title
`description`	string	no	Sub-issue description
`assigneeId`	string	no	Assign sub-issue to user UUID
`labelIds`	`string[]`	no	Initial labels
`estimate`	number	no	Story point estimate

Output ports - success (LinearIssue) / fail ({ error: string })

Comment node

`linear.comment.create`

Creates a comment on a Linear issue. Supports Markdown formatting. Used by the SDLC templates to post progress updates, summaries, and failure messages back to the issue thread.

Input ports

Port	Type	Required	Description
`issueId`	string	yes	Linear issue UUID
`body`	string	yes	Comment body (Markdown)
`organizationId`	string	no	Override org context

Output ports - success ({ id: string; createdAt: string }) / fail ({ error: string })

Text utility

`linear.text.strip_mentions`

Strips Linear @mention syntax from a string. Returns the cleaned text and a list of mentioned user IDs. Use this before passing comment bodies to agents or LLM nodes - the raw mention syntax (@[User Name](mention://user/UUID)) is rarely what you want in a prompt.

Input ports

Port	Type	Required
`text`	string	yes

Output ports

Port	Field	Description
`success`	`{ strippedText: string; mentionedUserIds: string[] }`	Cleaned text + extracted user IDs

Example: SDLC stage open/close pattern

The following YAML excerpt shows how the SDLC v2 template wires acknowledge and close around a dispatch:

steps:
  - id: acknowledge
    type: action
    nodeId: linear.agent_session.acknowledge
    config:
      sessionId: "{{ $trigger.data.sessionId }}"

  - id: dispatch
    type: action
    nodeId: agent.invoke
    config:
      stageEvent: "{{ $trigger.data }}"

  - id: close
    type: action
    nodeId: linear.agent_session.close
    config:
      sessionId: "{{ $trigger.data.sessionId }}"
      reason: "Development stage dispatched. I'll continue in the session thread."

Linear Trigger Nodes - linear.agent_session.created, linear.issue.assigned, etc.
Condition Nodes - linear.issue.predicate, linear.comment.body_prefix_switch
Agent Action Nodes - agent.dispatch_stage, agent.invoke
Linear Integration - OAuth setup, required scopes, AgentSession documentation
Linear AgentSession Lifecycle - the full session state machine and platform behavior

On this page