GthRunStats
Defined in: _worktrees/docs-release/gaunt-sloth/packages/core/src/core/types.ts:81
GS2-16 — per-run analytics harvested from a finished agent turn, threaded into the opt-in
history recorder (recordSessionSafe) so gth insights reports real numbers instead of
zeros. All fields are best-effort: token counts are only present when the provider actually
reported usage_metadata (otherwise omitted so the recorder stores NULL and the insights
formatter suppresses the misleading 0), and tools lists the names of tools invoked during
the run (deduplicated, order-insensitive). There is no costUsd — cost requires a reliable
price table this project does not carry, so it is deliberately never invented here.
Extended by
Section titled “Extended by”Properties
Section titled “Properties”tokensInput?
Section titled “tokensInput?”
optionaltokensInput?:number
Defined in: _worktrees/docs-release/gaunt-sloth/packages/core/src/core/types.ts:83
Total prompt/input tokens across the run’s LLM calls, when the provider reported usage.
tokensOutput?
Section titled “tokensOutput?”
optionaltokensOutput?:number
Defined in: _worktrees/docs-release/gaunt-sloth/packages/core/src/core/types.ts:85
Total completion/output tokens across the run’s LLM calls, when the provider reported usage.
toolResults?
Section titled “toolResults?”
optionaltoolResults?:GthToolResult[]
Defined in: _worktrees/docs-release/gaunt-sloth/packages/core/src/core/types.ts:95
BATCH-21 — one record per executed tool result (ToolMessage) observed during the run, in
arrival order and NOT deduplicated (a tool called twice yields two records), so gth eval’s
tool-RESULT assertions (must_error / tool_result_json_path) can grade what a tool
returned, not just that it was called. Optional (additive): producers that predate the field
simply omit it; runStats.js finalizeRunStats always sets it.
tools:
string[]
Defined in: _worktrees/docs-release/gaunt-sloth/packages/core/src/core/types.ts:87
Names of tools invoked during the run (deduplicated); empty when no tools were used.