ReadonlycwdThe temp working dir (inspect or clean it up).
ReadonlyexitReadonlyhooksThe hooks that fired during the run, each with its decision — parsed from
the CLI's hook_response stream events. Same capture requirement as
toolCalls (empty without the stream). Lets a test assert hook firing
honestly instead of via a marker file.
ReadonlymodelThe requests the model received, captured by the scripted mock — each with
its system prompt and messages, flattened to text. Lets a test assert
what actually reached the model (a SessionStart hook's injected context, a
slash command's expansion), not just that a hook fired. Harness tier
only: the mock sees the requests, so this is populated by runHarnessTest
(with or without transcript); the eval tier drives the real API, so its
modelRequests is always empty.
ReadonlyoutputThe agent's final answer text (the terminal result event), or "".
ReadonlystderrHook block messages and diagnostics land here.
ReadonlystdoutOptional ReadonlysubagentsSub-agent (Task) runs as nested traces, keyed by subagent_type. A
subagent runs its own session; CC tags its events with parent_tool_use_id
(= the Task tool call) so its tool calls are recovered into a sub-trace
here, lettng a test assert what the subagent DID (not just that Task fired).
Empty unless the stream was captured / the harness emits subagent events.
ReadonlytoolThe tools the agent invoked, each paired with its result — parsed from the
transcript. Empty unless the run captured the stream (transcript: true on
the harness tier; always on the eval tier). Lets a test assert on the
agent's actions (skills, MCP tools, subagents) instead of grepping stdout.
ReadonlyturnsNumber of model turns the agent took (mock turns served).
The observable record of ONE run — the unified shape produced by BOTH testing tiers:
runHarnessTest's result andrunEval'smeasurectx (eval.ts) both satisfy it. That's what lets the bare predicates inharness-assert.ts(usedTool/skillResolved/toolCount/toolUsedWith/hookFired/outputContains) run over either, with the testing helpers asserting and eval measuring over the same vocabulary.