Trilo as a task management MCP

Install

Endpoint: https://api.trilo.chat/mcp. Auth is OAuth 2.1 with Dynamic Client Registration — the client opens a browser once; no PAT required.

{
  "mcpServers": {
    "trilo": {
      "url": "https://api.trilo.chat/mcp"
    }
  }
}

claude mcp add --transport http trilo https://api.trilo.chat/mcp

PAT fallback: create a token in Profile → API Tokens with mcp:read / mcp:write, then pass it as a Bearer header if your client does not support OAuth yet.

Task kit (start here)

Trilo exposes the full workspace, but agents usually need this subset first:

• list_tasksSearch and filter tasks (use assignedToMe for your plate)
• get_taskFull task details + activity by UUID or custom ID (MVP-123)
• create_taskCreate a task — project optional (lands as TASK-N)
• update_taskChange status, priority, assignees, due date
• list_task_commentsRead comments on a task
• add_task_commentPost a comment after implementing something
• get_current_userWho you are + workspace membership
• list_membersResolve teammate IDs for assignment
• send_messagePost into a project or chat thread

Copy-paste prompts

List my open tasks in Trilo with assignedToMe=true, sorted by priority.

Get task MVP-XXX from Trilo, implement it, then add a comment summarizing what changed and mark it waiting_for_review.

Create a Trilo task titled "…" with a short description of what we just discussed. No project needed.

Trilo MCP vs Linear MCP

Linear's MCP is excellent at issue CRUD. Trilo matches that for tasks, then gives the agent the rest of the work surface: the chat where the task was discussed, the page that holds the spec, the knowledge graph of related entities, and send_message so status updates land where humans actually look.

Choose Linear if you want a dedicated issue tracker. Choose Trilo if you want the agent to own the ticket and the thread it lives in. See also Trilo vs Linear.