Linear

Write the returned comment.id to .workpad-id so subsequent syncs can update.

Subsequent syncs — read .workpad-id, update in place:

mutation UpdateComment($id: String!, $body: String!) {
  commentUpdate(id: $id, input: { body: $body }) { success }
}

Query an issue

The orchestrator injects issue context (identifier, title, description, state, labels, URL) into your prompt at startup. You usually do not need to re-read.

When you do, use the narrowest lookup for what you have:

# By ticket key (e.g. MT-686)
query($key: String!) {
  issue(id: $key) {
    id identifier title url description
    state { id name type }
    project { id name }
  }
}

For comments and attachments:

query($id: String!) {
  issue(id: $id) {
    comments(first: 50) { nodes { id body resolvedAt user { name } createdAt } }
    attachments(first: 20) { nodes { url title sourceType } }
  }
}

Interpretation note:

• Linear comment objects expose resolvedAt, not resolved.
• Active/unresolved comments have resolvedAt: null.
• Resolved comments have a non-null resolvedAt timestamp.

State transitions

Fetch team states first, then move with the exact stateId:

query($id: String!) {
  issue(id: $id) {
    team { states { nodes { id name } } }
  }
}

mutation($id: String!, $stateId: String!) {
  issueUpdate(id: $id, input: { stateId: $stateId }) {
    success
    issue { state { name } }
  }
}

Attach a PR or URL

# GitHub PR (preferred for PRs)
mutation($issueId: String!, $url: String!, $title: String) {
  attachmentLinkGitHubPR(issueId: $issueId, url: $url, title: $title, linkKind: links) {
    success
  }
}

# Plain URL
mutation($issueId: String!, $url: String!, $title: String) {
  attachmentLinkURL(issueId: $issueId, url: $url, title: $title) {
    success
  }
}

File upload

Three steps:

1. Get upload URL:

mutation($filename: String!, $contentType: String!, $size: Int!) {
  fileUpload(filename: $filename, contentType: $contentType, size: $size, makePublic: true) {
    success
    uploadFile { uploadUrl assetUrl headers { key value } }
  }
}

2. PUT file bytes to uploadUrl with the returned headers (use curl).
3. Embed assetUrl in comments/workpad as ![description](url).

Issue creation

Resolve project slug to IDs first:

query($slug: String!) {
  projects(filter: { slugId: { eq: $slug } }) {
    nodes { id teams { nodes { id key states { nodes { id name } } } } }
  }
}

Then create:

mutation($input: IssueCreateInput!) {
  issueCreate(input: $input) {
    success
    issue { identifier url }
  }
}

$input fields: title, teamId, projectId, and optionally description, priority (0–4), stateId. For relations, follow up with:

mutation($input: IssueRelationCreateInput!) {
  issueRelationCreate(input: $input) { success }
}

Input: issueId, relatedIssueId, type (blocks or related).

Rules

• No introspection. Never use __type or __schema queries. They return the entire Linear schema (~200K chars) and waste the context window. Every pattern you need is documented above.
• Keep queries narrowly scoped — ask only for fields you need.
• Sync the workpad at milestones, not after every change.
• For state transitions, always fetch team states first — never hardcode state IDs.
• Prefer attachmentLinkGitHubPR over generic URL attachment for GitHub PRs.