Records stored only a run's token/turn totals, so an arm's turn cost — the dominant driver of cache-read tokens — could not be diagnosed from the store. Retain the ordered transcript of tool invocations (the exact shell commands, MCP calls, and built-in tools the run made) on every scored record, absent only for a hung run that produced no transcript. The canonical TranscriptEntry shape lives on the record (result.ts); the isolation audit's ToolUse now aliases it so the persisted and audited shapes cannot drift.
3.9 KiB
3.9 KiB