rehydrate

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Workflow Rehydrate Skill

Overview

概述

Restore full workflow awareness that survives context auto-summarization without starting a new session.

This skill owns the read/restore side of workflow state — reading current state, reconciling it against git reality, and verifying whether a workflow exists at all. For the write side (init, update, transition, capturing a handoff), see

@skills/checkpoint/SKILL.md

恢复完整的工作流感知能力，使其在上下文自动总结后仍能保留，无需启动新会话。

本技能负责工作流状态的读取/恢复侧——读取当前状态、与git实际状态协调，并验证工作流是否存在。至于写入侧（初始化、更新、转换、捕获交接），请参阅

@skills/checkpoint/SKILL.md

。

Triggers

触发条件

Activate this skill when:

Restoring context after summarization (
```
rehydrate <featureId>
```
)
The agent has drifted after context compaction (stopped emitting events or using tools proactively)
Returning to a workflow after a break

在以下场景激活本技能：

总结后恢复上下文（
```
rehydrate <featureId>
```
）
Agent在上下文压缩后出现偏离（停止发送事件或主动使用工具）
中断后回到工作流

State Location

状态位置

Workflow state lives in the MCP event store, not the filesystem. Use

exarchos:exarchos_workflow get

to read state and

exarchos_view pipeline

to discover active workflows. Do not scan

~/.claude/workflow-state/*.state.json

— that path is legacy and may be stale or empty.

工作流状态存储在MCP事件存储中，而非文件系统。使用

exarchos:exarchos_workflow get

读取状态，使用

exarchos_view pipeline

发现活跃工作流。请勿扫描

~/.claude/workflow-state/*.state.json

——该路径属于遗留路径，内容可能过时或为空。

Source of Truth — does this workflow exist?

真实性来源——该工作流是否存在？

Workflow state lives in two surfaces, and conflating them causes wrong "untracked" conclusions:

The SQLite event store (
```
events
```
+ projected
```
workflow_state
```
+
```
streams
```
) — the authoritative record of whether a workflow exists. This is what
```
rehydrate
```
/
```
get
```
read.
<featureId>.state.json
files (under the state dir) — a secondary "planner's stamp" that carries plan-state facts the event projection cannot derive (review status, declared task list, dimension findings). It may be absent for a tracked workflow (CLI tools, tests, in-flight workflows before the first
```
update
```
) and is not an existence signal.

Canonical existence check: use the rehydrate envelope's _meta.workflowExists
(

true

false

), or equivalently a non-empty

data.workflowState.featureId

. A cold probe of a never-

init

'd featureId returns

success: true

with an empty initial document and

_meta.workflowExists: false

— and is side-effect-free (it emits no

workflow.rehydrated

event). Never infer existence from the presence or absence of a
.state.json
file on disk.

工作流状态存在于两个层面，混淆两者会导致错误的“未跟踪”结论：

SQLite事件存储（
```
events
```
+ 投影的
```
workflow_state
```
+
```
streams
```
）——这是工作流是否存在的权威记录，也是
```
rehydrate
```
/
```
get
```
读取的来源。
<featureId>.state.json
文件（位于状态目录下）——这是一个次要的“规划器标记”，承载事件投影无法推导的计划状态信息（评审状态、声明的任务列表、维度发现）。对于已跟踪的工作流，它可能不存在（CLI工具、测试、首次
```
update
```
前的进行中工作流），因此不能作为存在的信号。

标准存在性检查：使用rehydrate信封中的**

_meta.workflowExists

**（

true

false

），或等效的非空

data.workflowState.featureId

。对从未初始化的featureId进行冷探测会返回

success: true

，同时包含空的初始文档和

_meta.workflowExists: false

——且无副作用（不会触发

workflow.rehydrated

事件）。切勿根据磁盘上
.state.json
文件的存在与否推断工作流是否存在。

Rehydrate the Workflow

恢复工作流

Use

exarchos:exarchos_workflow

with

action: "rehydrate"

and

featureId: "<id>"

— it returns an envelope containing the canonical rehydration document (

workflowState

taskProgress

artifacts

blockers

, phase playbook, next actions) in a single call. No multi-step

get fields=[...]

composition is needed.

If the featureId is unknown or the user hasn't named one, fall back to

exarchos_view pipeline

to list active workflows and ask which to rehydrate, then re-invoke

rehydrate

with the selected

featureId

. The pipeline view is repo-scoped by default — it lists only the caller's repo, so a workflow started in another repo won't appear. Every response reports

unscopedTotal

(the pre-scope count); when

unscopedTotal

exceeds

page.total

, the hidden rows live in other repos (or are legacy rows without recorded identity) — re-query with

scope: "all"

(or an explicit

repoRoot

) to reveal them.

使用

exarchos:exarchos_workflow

，设置

action: "rehydrate"

和

featureId: "<id>"

——它会返回一个信封，其中包含标准的恢复文档（

workflowState

、

taskProgress

、

artifacts

、

blockers

、阶段手册、下一步操作），一次调用即可完成。无需多步

get fields=[...]

组合。

如果featureId未知或用户未指定，可回退使用

exarchos_view pipeline

列出活跃工作流，询问用户要恢复哪个，然后使用选定的

featureId

重新调用

rehydrate

。管线视图默认按仓库范围筛选——仅列出调用者所在仓库的工作流，因此在其他仓库启动的工作流不会显示。每个响应都会报告

unscopedTotal

（未筛选前的总数）；当

unscopedTotal

超过

page.total

时，隐藏的条目存在于其他仓库（或属于未记录身份的遗留条目）——使用

scope: "all"

（或显式指定

repoRoot

）重新查询即可显示这些条目。

Read State (targeted)

定向读取状态

For a targeted read rather than a full rehydration, use

exarchos:exarchos_workflow

with

action: "get"

and

featureId

Full state: Call with just
```
featureId
```
Specific field: Add
```
query
```
for dot-path lookup (e.g.,
```
query: "phase"
```
,
```
query: "tasks"
```
)
Multiple fields: Add
```
fields
```
array for projection (e.g.,
```
fields: ["phase", "featureId", "tasks"]
```
)

Field projection via

fields

returns only the requested top-level keys, reducing token cost.

若只需定向读取而非完整恢复，可使用

exarchos:exarchos_workflow

，设置

action: "get"

和

featureId

：

完整状态：仅传入
```
featureId
```
调用
特定字段：添加
```
query
```
进行点路径查找（例如
```
query: "phase"
```
、
```
query: "tasks"
```
）
多个字段：添加
```
fields
```
数组进行投影（例如
```
fields: ["phase", "featureId", "tasks"]
```
）

通过

fields

进行字段投影仅返回请求的顶级键，可降低token消耗。

Get Summary

获取摘要

For context restoration after summarization, prefer

action: "rehydrate"

(single-call, includes the phase playbook and next actions). For a minimal read,

action: "get"

with

featureId

outputs a summary suitable for rebuilding orchestrator context.

对于总结后的上下文恢复，优先使用

action: "rehydrate"

（单次调用，包含阶段手册和下一步操作）。若只需最小化读取，使用

action: "get"

并传入

featureId

，输出的摘要适用于重建编排器上下文。

Output Format

输出格式

Render the returned document as compact behavioral context (the same shape as post-compaction context) so the agent refreshes its awareness in one pass:

markdown

undefined

将返回的文档渲染为紧凑的行为上下文（与压缩后的上下文格式一致），以便Agent一次性刷新感知：

markdown

undefined

Workflow Rehydrated: <featureId>

工作流已恢复: <featureId>

Phase: <phase> | Type: <workflowType>

阶段: <phase> | 类型: <workflowType>

House Rules (apply every action this turn forward)

规则（本次操作的所有动作均需遵守）

Skill: <phasePlaybook.skillRef or "(no playbook for this phase)"> Tools: <phasePlaybook.tools rendered as bullets> Required model-emitted events: <phasePlaybook.events rendered as bullets — e.g.

task.progressed

phase.advanced

> Auto-emitted events (runtime fires these): <phasePlaybook.autoEmittedEvents rendered as bullets> Transition: <phasePlaybook.transitionCriteria> | Guard: <phasePlaybook.guardPrerequisites> Validation scripts: <phasePlaybook.validationScripts joined>

技能: <phasePlaybook.skillRef 或 "(本阶段无手册)"> 工具: <phasePlaybook.tools 以项目符号形式呈现> 模型需触发的事件: <phasePlaybook.events 以项目符号形式呈现 —— 例如

task.progressed

phase.advanced

> 自动触发的事件（运行时触发）: <phasePlaybook.autoEmittedEvents 以项目符号形式呈现> 转换条件: <phasePlaybook.transitionCriteria> | 前置检查: <phasePlaybook.guardPrerequisites> 验证脚本: <phasePlaybook.validationScripts 拼接显示>

Event Emission Hints

事件触发提示

<_eventHints.missing rendered as bullets, or "(none — phase machinery satisfied)">

<_eventHints.missing 以项目符号形式呈现，或 "(无 —— 阶段机制已满足)">

Task Progress

任务进度

<任务表格>

Artifacts

工件

Design: <path or "not created">
Plan: <path or "not created">
PR: <url or "not created">

设计: <路径或 "未创建">
计划: <路径或 "未创建">
PR: <链接或 "未创建">

Next Action

下一步操作

<suggested action, from the envelope's

next_actions

Discipline reminder: every task transition this turn forward MUST land on the workflow event stream via
exarchos_event.append
or
delegate
subagent emission. Direct
Edit
/
Bash
/
git
actions on task branches without corresponding events will desync the workflow tracker (see RCA
docs/rca/2026-05-08-rehydrate-behavioral-gap.md
).


Keep the output minimal — only essential state and behavioral guidance; full details stay in files, not the conversation.

<建议动作，来自信封中的

next_actions

纪律提醒: 本次操作的所有任务转换必须通过
exarchos_event.append
或
delegate
子Agent触发，写入工作流事件流。若直接对任务分支执行
Edit
/
Bash
/
git
动作而未触发对应事件，会导致工作流跟踪器不同步（请参阅RCA文档
docs/rca/2026-05-08-rehydrate-behavioral-gap.md
）。


保持输出简洁——仅包含必要的状态和行为指导；完整细节保留在文件中，而非对话里。

Reconcile State

协调状态

To verify state matches git reality, run

rehydrate <featureId>

— the rehydration projection folds events newer than the last snapshot and surfaces drift in the returned envelope. For deeper manual verification, run the reconciliation script:

typescript

exarchos_orchestrate({
  action: "reconcile_state",
  stateFile: "<state-file>",
  repoRoot: "<repo-root>"
})

On
passed: true
: State is consistent. On
passed: false
: Discrepancies found — review output and resolve via

exarchos:exarchos_workflow

with

action: "update"

(see

@skills/checkpoint/SKILL.md

要验证状态与git实际状态匹配，运行

rehydrate <featureId>

——恢复投影会合并最后一次快照之后的事件，并在返回的信封中显示偏差。如需更深入的手动验证，运行协调脚本：

typescript

exarchos_orchestrate({
  action: "reconcile_state",
  stateFile: "<state-file>",
  repoRoot: "<repo-root>"
})

当
passed: true
时：状态一致。当
passed: false
时：发现差异——查看输出并通过

exarchos:exarchos_workflow

设置

action: "update"

解决（请参阅

@skills/checkpoint/SKILL.md

）。

Best Practices

最佳实践

Reconcile on resume - Always verify state matches git state before acting
Read state, don't remember - After summarization, read from the event store, not memory
Single-call fetch - One
```
rehydrate
```
call returns the full canonical document; avoid multi-step reads
Existence via
_meta.workflowExists
- Never infer existence from a
```
.state.json
```
file on disk

恢复时进行协调 - 行动前始终验证状态与git状态匹配
读取状态，而非依赖记忆 - 总结后从事件存储读取，而非依赖内存
单次调用获取 - 一次
```
rehydrate
```
调用即可返回完整的标准文档；避免多步读取
通过
_meta.workflowExists
判断存在性 - 切勿根据磁盘上
```
.state.json
```
文件的存在与否推断工作流是否存在

Troubleshooting

故障排除

State Desync

状态不同步

If workflow state doesn't match git reality:

Run
```
rehydrate <featureId>
```
— the rehydration projection folds in events newer than the last snapshot
If manual check still needed: compare the rehydration document's
```
workflowState
```
/
```
artifacts
```
with
```
git log
```
and branch state

Update state via

exarchos:exarchos_workflow

with

action: "update"

to match git truth (see

@skills/checkpoint/SKILL.md

)

若工作流状态与git实际状态不匹配：

运行
```
rehydrate <featureId>
```
——恢复投影会合并最后一次快照之后的事件
若仍需手动检查：将恢复文档中的
```
workflowState
```
/
```
artifacts
```
与
```
git log
```
和分支状态对比
通过
```
exarchos:exarchos_workflow
```
设置
```
action: "update"
```
更新状态，使其与git实际状态匹配（请参阅
```
@skills/checkpoint/SKILL.md
```
）

Resume Finds Stale State

恢复时发现过时状态

If state references branches or worktrees that no longer exist:

Run
```
rehydrate <featureId>
```
— the rehydration document surfaces stale references
Compare against
```
git branch -a
```
/
```
git worktree list
```
to identify drift
Update via
```
exarchos_workflow update
```
to match git truth

若状态引用的分支或工作树已不存在：

运行
```
rehydrate <featureId>
```
——恢复文档会显示过时的引用
与
```
git branch -a
```
/
```
git worktree list
```
对比，识别偏差
通过
```
exarchos_workflow update
```
更新状态，使其与git实际状态匹配

Multiple Active Workflows

存在多个活跃工作流

If multiple workflow state files exist:

The system uses the most recently updated active (non-completed) workflow

Use

exarchos:exarchos_workflow

with

action: "cancel"

and

dryRun: true

on stale workflows to preview cleanup

Cancel stale workflows before starting new ones

若存在多个工作流状态文件：

系统会使用最近更新的活跃（未完成）工作流

对过时工作流使用

exarchos:exarchos_workflow

设置

action: "cancel"

和

dryRun: true

，预览清理操作

启动新工作流前取消过时工作流

Example Workflow

示例工作流

Resume after context loss: Use

exarchos:exarchos_workflow

with

action: "rehydrate"

and

featureId: "user-authentication"

to get context restoration output.

Check state: Use

exarchos:exarchos_workflow

with

action: "get"

and

featureId: "user-authentication"

Verify existence: Read
```
_meta.workflowExists
```
from the rehydrate envelope — if
```
false
```
, the feature was never started as a workflow, so report that rather than declaring it "untracked" from a filesystem check.

上下文丢失后恢复：使用
```
exarchos:exarchos_workflow
```
，设置
```
action: "rehydrate"
```
和
```
featureId: "user-authentication"
```
，获取上下文恢复输出。

检查状态：使用

exarchos:exarchos_workflow

，设置

action: "get"

和

featureId: "user-authentication"

。

验证存在性：从恢复信封中读取
```
_meta.workflowExists
```
——若为
```
false
```
，说明该功能从未作为工作流启动，因此应报告此情况，而非通过文件系统检查判定为“未跟踪”。