fleet_claim
Pack: fleet · Endpoint: https://gateway.pipeworx.io/fleet/mcp
Claim a task so no other worker takes it. Atomic: if two workers race for the same task exactly one wins and the loser gets claimed=false (not an error) and should move on. A claim carries a TTL; let it lapse without a heartbeat and the task auto-releases, so a dead session can never strand work.
Parameters
| Name | Type | Required | Description |
|---|---|---|---|
task_id | number | yes | The task to claim. |
claimed_by | string | yes | The claiming worker, as ‘session-name@host (lane)’. Host identity matters: it makes a dead machine a visible fact. |
ttl_seconds | number | no | Claim lifetime before auto-release, max ${MAX_TTL_SECONDS}. Default ${DEFAULT_TTL_SECONDS}. Extend with fleet_heartbeat rather than asking for a long one up front. |
Example call
Arguments
{
"task_id": 1,
"claimed_by": "warren@laptop (backend)"
}
curl
curl -X POST https://gateway.pipeworx.io/fleet/mcp \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"fleet_claim","arguments":{"task_id":1,"claimed_by":"warren@laptop (backend)"}}}'
TypeScript (@pipeworx/sdk)
import { Pipeworx } from '@pipeworx/sdk';
const pipeworx = new Pipeworx();
const result = await pipeworx.call('fleet_claim', {
"task_id": 1,
"claimed_by": "warren@laptop (backend)"
});
More examples
{
"task_id": 1,
"claimed_by": "warren@mini-1 (backend)",
"ttl_seconds": 3600
}
Connect
Add this to your MCP client config, or use one-click install buttons:
{
"mcpServers": {
"fleet": {
"url": "https://gateway.pipeworx.io/fleet/mcp"
}
}
}
See Getting Started for client-specific install steps.