Cost and usage tools
Cost of work per task and source, account pool records and the usage of tools per role. This group has 4 tools, 3 of them available only with spec-editor Pro.
| Tool | Edition | Summary |
|---|---|---|
economy_get |
Pro | Cost aggregates over task records: the cost of every plan, task, run and subagent call, rolled up bottom-up so that a parent is the sum of its children; the… |
kpi_values |
Pro | KPI values written by the hourly evaluation: per KPI the value, window, computed_at, 30-day daily history and trend, with the target and whether it deviates. |
pool_records_get |
Pro | Every account pool record of the llm.models setting: account id, provider, tariff, rotation state (free, parked or unusable), the last-polled utilisation and… |
tool_usage |
Open core | Calls per role x tool from mcp/tool_called events and coder/tester tool_use records, with p50/p95 latency, read-vs-write share, top elements read, and… |
economy_get
Pro tool. Available with the spec-editor Pro plugin.
Cost aggregates over task records: the cost of every plan, task, run and subagent call, rolled up bottom-up so that a parent is the sum of its children; the attempts made per task with model, effort and outcome; the 7-day delivery tiles (cost to confirmed, autonomy); and the accuracy of the difficulty classifier over the same window. Read-only.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
plan_id |
string | no | Narrow plans to one plan; omit for all |
project_path |
string | yes | Path to spec-editor project (directory with methodology.yaml) |
window_days |
integer | no | Tile window in days (default 7) |
kpi_values
Pro tool. Available with the spec-editor Pro plugin.
KPI values written by the hourly evaluation: per KPI the value, window, computed_at, 30-day daily history and trend, with the target and whether it deviates.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
kpi_id |
string | no | Only this KPI |
project_path |
string | no | Path to spec-editor project |
pool_records_get
Pro tool. Available with the spec-editor Pro plugin.
Every account pool record of the llm.models setting: account id, provider, tariff, rotation state (free, parked or unusable), the last-polled utilisation and reset time of the five-hour and seven-day limit windows, the reason a record was parked, and the account’s own spend over the window plus the pool’s total. Also lists the record each role is using now and the parked or unusable records that are waiting. Identifiers and metadata only: never a token, a key or a credential path. Read-only.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
project_path |
string | yes | Path to spec-editor project (directory with methodology.yaml) |
window_days |
integer | no | Spend window in days (default 7) |
tool_usage
Calls per role x tool from mcp/tool_called events and coder/tester tool_use records, with p50/p95 latency, read-vs-write share, top elements read, and tokens/$ per role (window) and per task. Unavailable figures are null. Read-only.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
project_path |
string | yes | Path to spec-editor project (directory with methodology.yaml) |
since |
string | yes | Window start, ISO-8601 |
until |
string | no | Window end, ISO-8601 (exclusive) |