Harness² Demo — runs/hermes-test

domain: accounting · tasks ok: 2/2 · failed: 0 · executions: 28 · editor_calls: 32

taskstatus
ledger-1ok
ledger-2ok

runs

A.md

B.md

EXPERIENCE.md

INDEX.md

LongTermMEMORY.md

README.md

adoption.json

agg1.md

agg2.md

answer.txt

base1.md

base2.md

c1.md

c2.md

dom1.md

dom2.md

edits.md

guardrails.md

harness.diff

inv1.md

inv2.md

ledger.csv

produced.md

rollout.md

selection.json

summarize.py

summary.json

summary.lock

systemprompt.md

task.md

tool_guidance.md

library

sources

刷新方式:由 HTTP 服务实时读取目录,新生成的文件重新打开本页即可看到。