orchestrate-omp

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Orchestrate on OMP

基于OMP进行编排

Stay available to the user while substantive work runs behind the scenes. The root coordinator owns the plan, user interaction, approvals, scope, integration, and final verification.

在后台执行重要工作期间，需保持对用户的响应性。根协调者负责计划制定、用户交互、审批、范围界定、集成以及最终验证。

1. Frame

1. 框架搭建

State the done predicate and the final artifact or decision.
Split only independent discovery, specialist review, or implementation with clear ownership. Keep tightly coupled work in one session.
Assign distinct primary ownership and name shared contracts before dispatch. Do not duplicate investigations or serialize a useful batch merely to avoid possible file overlap.
Keep external side effects and approval decisions with the root.

明确完成判定条件以及最终产出物或决策结果。
仅拆分具有明确归属的独立探索、专家评审或实施任务。紧密关联的工作需保留在同一会话中。
在分派任务前，分配明确的主要归属权并定义共享契约。不要为了避免可能的文件重叠而重复调查或序列化有效批次任务。
外部副作用及审批决策由根协调者负责。

2. Choose Agents

2. 选择代理

Use the exact names exposed by the live

task

tool; its inventory is authoritative. For the current OMP roster:

Work	Agent	Effort when exposed
Narrow read-only codebase discovery	`scout`	`lo`
Routine scoped implementation	Omit `agent` for the default worker	`med`
Difficult or ambiguous implementation	Omit `agent` for the default worker	`hi`
Mechanical edits or data collection	`sonic`	`lo`
UI/UX implementation or review	`designer`	`med` or `hi`
Code-quality verdict	`reviewer`	`med` or `hi`
Security verdict	`security-reviewer`	`hi`
External library or API research	`librarian`	`lo` or `med`

When the live task schema exposes

effort

, match it to the assignment:

lo

for narrow questions,

med

for routine work, and

hi

when ambiguity or consequence justifies it. Otherwise rely on the selected agent's configured model role; do not pass an unavailable field.

Honor read-only and blocking markers in the live inventory. Never pass the default worker's name explicitly.

使用实时

task

工具暴露的准确名称；其清单具有权威性。当前OMP代理清单如下：

工作内容	Agent	暴露时的工作强度
限定范围的只读代码库探索	`scout`	`lo`
常规限定范围实施	默认工作者可省略 `agent` 字段	`med`
复杂或模糊场景下的实施	默认工作者可省略 `agent` 字段	`hi`
机械编辑或数据收集	`sonic`	`lo`
UI/UX实施或评审	`designer`	`med` 或 `hi`
代码质量评估	`reviewer`	`med` 或 `hi`
安全评估	`security-reviewer`	`hi`
外部库或API调研	`librarian`	`lo` 或 `med`

当实时任务模式暴露

effort

字段时，需将其与任务匹配：

lo

适用于限定范围的问题，

med

适用于常规工作，当存在模糊性或严重后果时使用

hi

。否则依赖所选代理的配置模型角色；不要传递不存在的字段。

需遵守实时清单中的只读和阻塞标记。切勿显式传递默认工作者的名称。

3. Dispatch

3. 任务分派

When the live schema exposes
```
tasks[]
```
, put independent participants in one batch so they start together and structure shared
```
context
```
as
```
Goal
```
,
```
Constraints
```
, and
```
Contract
```
. Otherwise start one participant per task call without passing
```
context
```
.
Give every item a stable CamelCase
```
name
```
of at most 32 characters, the most specific specialist
```
agent
```
when one fits, and a complete standalone
```
task
```
. Omit
```
agent
```
only for the default worker. Add
```
effort
```
only when the live schema exposes it.
Structure each assignment as
```
Target
```
,
```
Change
```
, and
```
Acceptance
```
. Include essential repository state, decisions, restrictions, and dependencies because children start without conversation history.
Give every writer an exact writable scope. When the live schema exposes
```
isolated
```
, use
```
isolated: true
```
for concurrent writers that need separate workspaces, then inspect the returned metadata to learn whether changes were applied or retained as artifacts.
Tell ordinary workers:
```
Complete this assignment directly. Do not call task or spawn subagents.
```
Allow nested delegation only when the assignment explicitly makes that worker a coordinator.
Tell workers to skip project-wide formatting, linting, and test suites. The root runs shared validation once after integration.

Large context belongs in a local file referenced with

local://<path>

, not duplicated across assignments.

Use

outputSchema

only when the coordinator needs a machine-readable result. Set

schemaMode: "strict"

when invalid output must fail rather than return with a warning.

当实时模式暴露
```
tasks[]
```
字段时，将独立参与者放入同一批次以同步启动，并将共享
```
context
```
结构化为
```
Goal
```
（目标）、
```
Constraints
```
（约束）和
```
Contract
```
（契约）。否则每次任务调用启动一个参与者，无需传递
```
context
```
。
为每个任务项分配一个最多32字符的稳定驼峰式
```
name
```
，匹配最适合的专业
```
agent
```
，并提供完整独立的
```
task
```
内容。仅默认工作者可省略
```
agent
```
字段。仅当实时模式暴露
```
effort
```
字段时才添加该字段。
将每个任务分配结构化为
```
Target
```
（目标）、
```
Change
```
（变更）和
```
Acceptance
```
（验收标准）。需包含必要的仓库状态、决策、限制条件和依赖关系，因为子代理启动时无对话历史。
为每个写入者分配精确的可写入范围。当实时模式暴露
```
isolated
```
字段时，为需要独立工作区的并发写入者设置
```
isolated: true
```
，随后检查返回的元数据以了解变更是否已应用或作为产出物保留。
告知普通工作者：
```
直接完成此任务分配。请勿调用task或生成子代理。
```
仅当任务分配明确指定该工作者为协调者时，才允许嵌套委派。
告知工作者跳过项目级别的格式化、代码检查和测试套件。根协调者会在集成后统一执行共享验证。

大型上下文应放入通过

local://<path>

引用的本地文件中，不要在多个任务分配中重复。

仅当协调者需要机器可读结果时才使用

outputSchema

。当无效输出必须直接失败而非返回警告时，设置

schemaMode: "strict"

。

4. Coordinate

4. 任务协调

Results auto-deliver. Continue handling user messages and independent root work instead of polling continuously.
Siblings may exchange concise dependency updates through OMP IRC. Name expected dependencies in shared context so communication is purposeful rather than discovery by negotiation.
Use the live job-control surface only to wait for required work, inspect status, send a bounded correction, or cancel stale work.
Read complete results from
```
agent://<id>
```
and use
```
history://<id>
```
when a report is incomplete or suspicious.
A completed job means the child yielded successfully. It does not mean its artifact is accepted.

结果会自动交付。继续处理用户消息和根协调者的独立工作，无需持续轮询。
同级代理可通过OMP IRC交换简洁的依赖更新。在共享上下文中明确预期依赖，使沟通更具目的性而非通过协商探索。
仅在等待必要工作、检查状态、发送有限修正或取消过期工作时使用实时作业控制界面。
从
```
agent://<id>
```
读取完整结果，当报告不完整或存在疑点时使用
```
history://<id>
```
。
作业完成仅表示子代理成功提交结果，不代表其产出物已被接受。

5. Integrate

5. 结果集成

Drain every required participant or record the cancellation or gap.
Inspect each claimed artifact and reconcile conflicts against the declared ownership and contracts.
Run the affected validation and real user path from the root at the integrated head.
Return one concise result with agent identifiers, accepted findings or changes, verification evidence, and unresolved gaps.

Claim model or backend diversity only when returned metadata proves it.

收集所有必要参与者的结果，或记录取消情况与缺口。
检查每个提交的产出物，并根据已声明的归属权和契约解决冲突。
在集成后的最新版本上，从根协调者处运行受影响的验证及真实用户路径测试。
返回一份简洁结果，包含代理标识符、已接受的发现或变更、验证证据以及未解决的缺口。

仅当返回的元数据能够证明时，才可宣称模型或后端的多样性。