What it does
Close a run you opened with start_run, reporting what it spent and the evidence for what it claims. If tool_calls or the elapsed minutes exceed the task's budget, the run is recorded as over_budget (you can still finish — the overage is just on the record). Pass receipts so a human approving this work can check it instead of taking the summary's word for it: a commit SHA, a CI run URL, a test tally, a published doc. Posts a trace line to the task's activity.
Parameters
| Name | Type | Required | Description |
|---|---|---|---|
id | string | Required | Task id |
run_id | string | Required | The run id returned by start_run |
tool_calls | number | Optional | How many tool calls this run made |
tokens | number | null | Optional | Tokens spent, if you track them |
summary | string | null | Optional | Short note on what the run did |
outcome | string | Optional | Defaults to completed Allowed values completedfailed |
receipts | object[] | Optional | Evidence for what this run claims — what a reviewer would check instead of trusting the summary. Each needs a url or a detail (or both); a receipt with neither is just an assertion. |
receipts[].kind | string | Required | Allowed values commitcitestdocurl |
receipts[].label | string | Required | What this is, e.g. 'unit tests' or 'deploy' |
receipts[].url | string | null | Optional | Where to go and check it |
receipts[].detail | string | null | Optional | The checkable value itself — a SHA, '142 passed' |
Annotations
The hints the server sends with this tool. A host reads them to decide whether to ask you before a call.
- Read-only
readOnlyHint: false - Changes data in your Done account.
- Destructive
destructiveHint: false - Only adds or reads; nothing already there is removed or overwritten.
- Idempotent
idempotentHint: false - Each call can add something new, such as a comment, a run or a task.
- Open world
openWorldHint: false - Touches only your own Done account.
Which tokens can call it
A connection made with OAuth has full access. A token from Settings → Agent access → Extra MCP tokens can be narrowed, and a narrowed token never sees beyond its project or workspace.
- Full accessAllowed
- Read-only tokenRefused
- Single-project tokenRefused
- Single-workspace tokenAllowed
Try it
With Done connected, ask your assistant something like:
“When you're done, close the run with the commit and the test results as receipts.”
Connect Done to your assistant
One MCP endpoint per account, over OAuth or a token. Step-by-step setup for each client: