[!] TOPIC ARCHIVE // #AGENT EVALS
#Agent Evals
All 2 guides, operator dossiers, and signals tagged with #Agent Evals.
RADAR SIGNALRADAR SIGNAL
hidden tokens, long runs, repo memory
agent work showed its meter before the job even began: Systima measured harness token overhead, Long-Horizon-Terminal-Bench stretched terminal agents into hour-long runs, and capn-hook put repo memory on a token receipt.
exact calls, guarded writes, real tool tests
Agents are being pushed toward verifiable tool calls, pre-write service controls, and prompt tests that use the actual production harness.