A cost architecture: the main model plans, cheaper models execute, a stronger model reviews一套成本架构：主模型规划，便宜模型执行，更强模型把关
Frontier models are excellent planners and expensive typists. The cheapest way to run an agent is to stop paying flagship prices for the parts that do not need flagship judgment.顶级模型是优秀的规划者，也是昂贵的打字员。运行 Agent 最省钱的方式，就是不再为不需要顶级判断力的部分支付旗舰价格。
Any subtask that is mechanical and can be fully specified in writing goes to a cheap executor. On rote work the measured saving is roughly 2.3–2.7× versus running the same step inline on the main model — and for repetitive, template-shaped work the small model is often more accurate, because it follows the written contract instead of improvising.任何机械性、能用书面形式完整规格化的 subtask 都交给便宜的执行模型。在机械性工作上，相比在主模型上内联执行同一步骤，实测节省约 2.3–2.7×——而且对重复性、模板化的活儿，小模型常常更准，因为它照书面契约执行，而不是即兴发挥。
Judgment that cannot be specified goes the other way: up. A stronger advisor reviews the hardest calls. The main model keeps the three things it is genuinely needed for — planning, acceptance checks, and synthesis of results into the final answer.无法规格化的判断则走另一条路：向上。更强的 advisor 复核最难的判断。主模型保留三件真正需要它的事——规划、验收检查、把结果综合成最终答复。
A delegation is not a hint or a suggestion. It is a literal contract, written before the work starts.一次委托不是提示，也不是建议，而是一份字面意义上的契约，在工作开始之前就写好。
The contract states the GOAL, the INPUTS (explicit file paths — the executor sees nothing else), the OUTPUT format, and the forbidden side effects. Acceptance criteria are written before dispatch and are machine-checkable: the parent runs them itself on the executor's output, instead of trusting the executor's own report.契约写明 GOAL（目标）、INPUTS（输入：明确的文件路径——执行者看不到其他任何东西）、OUTPUT 格式，以及禁止的副作用。验收标准在派发之前写好，并且可被机器检查：由父方自己对着执行者的产出运行这些标准，而不是相信执行者自己的报告。
The executor runs in an isolated mini-session with file tools and bash, so a bad delegation cannot disturb the parent's state. Coordination is not free: the parent pays for every handoff, so tiny tasks stay inline, and one batched spec beats many small calls.执行者在带文件工具和 bash 的隔离迷你会话中运行，所以一次糟糕的委托也不会打扰父方的状态。协调并非免费：父方为每一次交接付费，因此微小任务留在内联执行，而一份批量规格胜过许多次小调用。
Advisor has two modes, and the difference between them is deliberately stark.Advisor 有两种模式，二者之间的差异刻意做得非常分明。
Consult mode sends a stronger model the full conversation context and asks for a plan or a correction when the agent is genuinely stuck. Verify mode is the opposite: an independent judge with no access to the conversation sees only the claim being made plus the evidence pasted next to it, and returns a verdict of PASS or FAIL. It is used before declaring any non-trivial deliverable done.咨询（consult）模式把完整对话上下文交给更强的模型，在 Agent 真正卡住时请它给方案或纠正。裁决（verify）模式恰恰相反：一位没有对话访问权的独立评审，只看被声称的结论以及旁边粘贴的证据，然后给出 PASS 或 FAIL 的判定。它用于宣布任何非平凡交付完成之前。
The no-context design is the point: a judge who saw the agent's reasoning would inherit the agent's own biases and simply agree. A judge who sees only the claim and the evidence has no choice but to check the evidence.不看上下文正是关键：看过 Agent 推理过程的评审会继承 Agent 自己的偏见，然后直接同意。只看结论和证据的评审别无选择，只能去核对证据。
Conductor Mode and the delegation architecture solve different problems, so they compose.Conductor Mode 与委托架构解决的是不同问题，因此可以组合使用。
Conductor picks the model for each user turn, deciding which lane a conversation travels in. Delegate and advisor reshape cost inside a single turn: the main model buys a cheap execution and an expensive review without leaving the conversation. All three are configurable per agent — the executor chain, the advisor model, and the cost caps — so each workroom tunes its own cost-quality trade.Conductor 为每一轮用户回合选模型，决定一段对话走哪条车道。Delegate 与 Advisor 则在一轮内部重塑成本：主模型买到一次便宜的执行和一次昂贵的复核，而无需离开会话。三者都按 agent 可配置——执行链、advisor 模型、成本上限——因此每个工作间可以调自己的成本-质量平衡。