Evidence library · family
Harness
15 tested rules in the 'harness' family, 15 held.
Last reviewed 2026-10-01
15 of 15 held niche Benchmarks, coverage & performance
What this family tests
Every rule the CAIN-42 test suites recorded under the family harness, across 1 evidence bundle. Each rule links to its own page with the recorded result and an in-browser check of the file it came from.
| ID | Rule | Bundle | Result |
|---|---|---|---|
| E35-HARNESS-plan | harness event plan is governed | E35 | held |
| E35-HARNESS-context_assembly | harness event context_assembly is governed | E35 | held |
| E35-HARNESS-memory | harness event memory is governed | E35 | held |
| E35-HARNESS-tool_selection | harness event tool_selection is governed | E35 | held |
| E35-HARNESS-tool_execution | harness event tool_execution is governed | E35 | held |
| E35-HARNESS-retry | harness event retry is governed | E35 | held |
| E35-HARNESS-loop | harness event loop is governed | E35 | held |
| E35-HARNESS-delegation | harness event delegation is governed | E35 | held |
| E35-HARNESS-checkpoint | harness event checkpoint is governed | E35 | held |
| E35-HARNESS-recovery | harness event recovery is governed | E35 | held |
| E35-HARNESS-state_persistence | harness event state_persistence is governed | E35 | held |
| E35-HARNESS-model_switching | harness event model_switching is governed | E35 | held |
| E35-HARNESS-browser_action | harness event browser_action is governed | E35 | held |
| E35-HARNESS-computer_use | harness event computer_use is governed | E35 | held |
| E35-HARNESS-code_execution | harness event code_execution is governed | E35 | held |
Other families in this niche
bench_family · bench_category · coverage · property · telemetry · fabric_component · radar · swe · regression · performance · corpus · end_to_end · completion · latency · f2t · false_completion · failure_to_test
Try CAIN-42 on your own agents
Create a free account and every new account starts with a 7-day trial of the full platform. Or try the sandbox first, with no account at all.
Create a free account → · Try the sandbox · See the whole ecosystem