workflow-state

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Workflow State Management Skill

工作流状态管理Skill

Overview

概述

Manage persistent workflow state that survives context auto-summarization.

State files store: task details, worktree locations, PR URLs, and review status.

管理可在上下文自动总结后保留的持久化工作流状态。

状态文件存储：任务详情、工作树位置、PR链接以及评审状态。

Triggers

触发条件

Activate this skill when:

Starting a new workflow (
```
/ideate
```
)
Transitioning between workflow phases
Restoring context after summarization (
```
/rehydrate
```
)
Saving progress for later continuation (
```
/checkpoint
```
)

在以下场景激活该Skill：

启动新工作流（
```
/ideate
```
）
工作流阶段间切换
总结后恢复上下文（
```
/rehydrate
```
）
保存进度以便后续继续（
```
/checkpoint
```
）

Phase Transitions

阶段切换

Valid transitions, guards, and prerequisites for all workflow types are documented in

references/phase-transitions.md

. CRITICAL: When a transition has a guard, send the prerequisite

updates

and

phase

in a single

set

call — updates apply before guards evaluate.

所有工作流类型的有效切换规则、防护条件及前置要求均记录在

references/phase-transitions.md

中。关键注意事项：当切换存在防护条件时，需在单次

set

调用中同时传入前置要求的

updates

和

phase

——更新操作会在防护条件评估前生效。

Schema Discovery

模式发现

Use

exarchos_workflow({ action: "describe", actions: ["set", "init", "get"] })

for parameter schemas and

exarchos_workflow({ action: "describe", playbook: "feature" })

for phase transitions, guards, and playbook guidance. For the lightweight oneshot variant (with its

implementing → synthesize|completed

choice state driven by

synthesisPolicy

), call

exarchos_workflow({ action: "describe", playbook: "oneshot" })

— oneshot is a first-class playbook alongside feature/debug/refactor. Use

exarchos_event({ action: "describe", eventTypes: ["workflow.transition", "task.completed"] })

for event data schemas.

使用

exarchos_workflow({ action: "describe", actions: ["set", "init", "get"] })

获取参数模式，使用

exarchos_workflow({ action: "describe", playbook: "feature" })

获取阶段切换规则、防护条件及工作流指南。对于轻量型一次性工作流变体（其

implementing → synthesize|completed

选择状态由

synthesisPolicy

驱动），请调用

exarchos_workflow({ action: "describe", playbook: "oneshot" })

——oneshot是与feature/debug/refactor同级的一等工作流。使用

exarchos_event({ action: "describe", eventTypes: ["workflow.transition", "task.completed"] })

获取事件数据模式。

State Location

状态存储位置

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

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_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

——该路径为遗留路径，内容可能过期或为空。

State Operations

状态操作

For full MCP tool signatures, error handling, and anti-patterns, see

references/mcp-tool-reference.md

关于MCP工具的完整签名、错误处理及反模式，请参考

references/mcp-tool-reference.md

。

Initialize State

初始化状态

At the start of

/ideate

, use

mcp__exarchos__exarchos_workflow

with

action: "init"

with:

```
featureId
```
: the workflow identifier (e.g.,
```
"user-authentication"
```
)

workflowType

: one of

"feature"

"debug"

"refactor"

"oneshot"

```
synthesisPolicy
```
(optional, oneshot only): one of
```
"always"
```
,
```
"never"
```
,
```
"on-request"
```
(default
```
"on-request"
```
) — silently ignored for non-oneshot types

This creates a new workflow state entry. The initial phase depends on

workflowType

```
feature
```
→ starts in
```
ideate
```
```
debug
```
→ starts in
```
triage
```
```
refactor
```
→ starts in
```
explore
```
```
oneshot
```
→ starts in
```
plan
```
(skips ideate entirely)

在

/ideate

启动时，使用

mcp__exarchos__exarchos_workflow

并设置

action: "init"

，同时传入：

```
featureId
```
：工作流标识符（示例：
```
"user-authentication"
```
）

workflowType

：可选值为

"feature"

、

"debug"

、

"refactor"

、

"oneshot"

```
synthesisPolicy
```
（可选，仅适用于oneshot）：可选值为
```
"always"
```
、
```
"never"
```
、
```
"on-request"
```
（默认值
```
"on-request"
```
）——非oneshot类型会自动忽略该参数

此操作将创建新的工作流状态条目。初始阶段取决于

workflowType

：

```
feature
```
→ 初始阶段为
```
ideate
```
```
debug
```
→ 初始阶段为
```
triage
```
```
refactor
```
→ 初始阶段为
```
explore
```
```
oneshot
```
→ 初始阶段为
```
plan
```
（直接跳过ideate阶段）

Workflow Types at a Glance

工作流类型概览

```
feature
```
— full
```
ideate → plan → delegate → review → synthesize
```
for real features with subagent dispatch and two-stage review

debug

—

triage → investigate → (thorough | hotfix)

for bug workflows with track selection

```
refactor
```
—
```
explore → brief → (polish | overhaul)
```
for code improvements, polish for small and overhaul for multi-task
```
oneshot
```
—
```
plan → implementing → (completed | synthesize)
```
for trivial changes; direct-commit by default with an opt-in PR path resolved via a choice-state guard driven by
```
synthesisPolicy
```
and the
```
synthesize.requested
```
event

See

@skills/oneshot-workflow/SKILL.md

for the lightweight variant's full prose, including the choice-state mechanics and

finalize_oneshot

trigger.

```
feature
```
— 完整的
```
ideate → plan → delegate → review → synthesize
```
流程，适用于包含子Agent调度和两阶段评审的正式功能开发
```
debug
```
—
```
triage → investigate → (thorough | hotfix)
```
流程，适用于包含路径选择的Bug修复工作流
```
refactor
```
—
```
explore → brief → (polish | overhaul)
```
流程，适用于代码优化，polish用于小型优化，overhaul用于多任务重构
```
oneshot
```
—
```
plan → implementing → (completed | synthesize)
```
流程，适用于微小变更；默认直接提交，可通过
```
synthesisPolicy
```
和
```
synthesize.requested
```
事件驱动的选择状态防护条件开启PR路径

关于轻量型变体的完整说明，包括选择状态机制和

finalize_oneshot

触发器，请参考

@skills/oneshot-workflow/SKILL.md

。

Read State

读取状态

Use

mcp__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.

使用

mcp__exarchos__exarchos_workflow

并设置

action: "get"

和

featureId

：

完整状态：仅传入
```
featureId
```
即可
特定字段：添加
```
query
```
参数进行点路径查询（示例：
```
query: "phase"
```
、
```
query: "tasks"
```
）
多个字段：添加
```
fields
```
数组进行投影查询（示例：
```
fields: ["phase", "featureId", "tasks"]
```
）

通过

fields

进行字段投影仅返回指定的顶层键，可降低Token消耗。

Update State

更新状态

Use

mcp__exarchos__exarchos_workflow

with

action: "set"

with

featureId

and either

updates

phase

, or both:

Update phase:
```
phase: "delegate"
```

Set artifact path:

updates: { "artifacts.design": "docs/designs/2026-01-05-feature.md" }

Mark task complete (by index):

updates: { "tasks[0].status": "complete", "tasks[0].completedAt": "<timestamp>" }

Add worktree:

updates: { "worktrees.wt-001": { "branch": "feature/001-types", "taskId": "001", "status": "active" } }

Phase + updates together:

phase: "delegate"

updates: { "artifacts.plan": "docs/plans/plan.md" }

Worktree status values:

'active' | 'merged' | 'removed'

使用

mcp__exarchos__exarchos_workflow

并设置

action: "set"

，同时传入

featureId

以及

updates

、

phase

或两者：

更新阶段：
```
phase: "delegate"
```

设置工件路径：

updates: { "artifacts.design": "docs/designs/2026-01-05-feature.md" }

标记任务完成（按索引）：

updates: { "tasks[0].status": "complete", "tasks[0].completedAt": "<timestamp>" }

添加工作树：

updates: { "worktrees.wt-001": { "branch": "feature/001-types", "taskId": "001", "status": "active" } }

同时更新阶段和字段：

phase: "delegate"

updates: { "artifacts.plan": "docs/plans/plan.md" }

工作树状态可选值：

'active' | 'merged' | 'removed'

Editing the

tasks

array

编辑

tasks

数组

The dot-path parser used by

set updates

recognizes only numeric array brackets (

tasks[0]

tasks[1]

, …). Keyed forms like

tasks[id=T-001]

are NOT supported and now throw an

INVALID_INPUT

error with a clear message — earlier versions silently wrote to a bogus top-level key, returning

success: true

while the actual task was untouched. Three patterns are supported:

Replace the whole array (use this when the plan is being revised wholesale; matches the issue #1003 contract):

typescript

exarchos_workflow({
  action: "set",
  featureId: "<id>",
  updates: { tasks: [
    { id: "T-001", title: "...", status: "pending" },
    { id: "T-002", title: "...", status: "pending" },
  ]},
})

Edit one task by its array index:

typescript

exarchos_workflow({
  action: "set",
  featureId: "<id>",
  updates: { "tasks[0].status": "complete", "tasks[0].completedAt": "<ts>" },
})

First read

tasks

(

action: "get", query: "tasks"

) to find the index of the task you want to edit, then set by that index.

Append a new task by writing to the next-free index. If the array currently has length

, write to

tasks[N]

typescript

// Suppose tasks already contains T-001 and T-002 (length 2). To append:
exarchos_workflow({
  action: "set",
  featureId: "<id>",
  updates: { "tasks[2]": { id: "T-003", title: "Follow-up", status: "pending" } },
})

The parser allows writing one slot past the current length (

MAX_ARRAY_GAP = 1

); writing further out (

tasks[5]

against a length-2 array) throws

INVALID_INPUT

. Read the current

tasks

length before appending.

set updates

使用的点路径解析器仅支持数字数组括号（

tasks[0]

、

tasks[1]

等）。

tasks[id=T-001]

这类键值形式不被支持，会直接抛出

INVALID_INPUT

错误并给出明确提示——早期版本会静默写入错误的顶层键，返回

success: true

但实际任务未被修改。支持以下三种模式：

替换整个数组（当计划需要全面修订时使用；符合issue #1003约定）：

typescript

exarchos_workflow({
  action: "set",
  featureId: "<id>",
  updates: { tasks: [
    { id: "T-001", title: "...", status: "pending" },
    { id: "T-002", title: "...", status: "pending" },
  ]},
})

按数组索引编辑单个任务：

typescript

exarchos_workflow({
  action: "set",
  featureId: "<id>",
  updates: { "tasks[0].status": "complete", "tasks[0].completedAt": "<ts>" },
})

先读取

tasks

（

action: "get", query: "tasks"

）找到目标任务的索引，再按该索引进行设置。

追加新任务：写入下一个可用索引。如果当前数组长度为

，则写入

tasks[N]

：

typescript

// 假设tasks已包含T-001和T-002（长度为2），追加新任务：
exarchos_workflow({
  action: "set",
  featureId: "<id>",
  updates: { "tasks[2]": { id: "T-003", title: "Follow-up", status: "pending" } },
})

解析器允许写入当前长度之后的一个位置（

MAX_ARRAY_GAP = 1

）；写入超出范围的位置（如长度为2的数组写入

tasks[5]

）会抛出

INVALID_INPUT

错误。追加前请先读取当前

tasks

的长度。

Get Summary

获取摘要

For context restoration after summarization, use

mcp__exarchos__exarchos_workflow

with

action: "get"

and

featureId

. This outputs a minimal summary suitable for rebuilding orchestrator context.

总结后恢复上下文时，使用

mcp__exarchos__exarchos_workflow

并设置

action: "get"

和

featureId

。此操作会输出用于重建编排器上下文的极简摘要。

Reconcile State

状态对账

To verify state matches git reality, the SessionStart hook automatically reconciles on resume. For 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.

为验证状态与git实际情况一致，SessionStart钩子会在恢复时自动执行对账操作。如需手动验证，运行对账脚本：

typescript

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

当
passed: true
时：状态一致。当
passed: false
时：发现不一致——请查看输出并解决问题。

Integration Points

集成点

When to Update State

状态更新时机

Event	State Update
`/ideate` starts	Create state file
Design saved	Set `artifacts.design` , phase = "plan"
Plan saved	Set `artifacts.plan` , populate tasks, phase = "plan-review"
Plan-review gaps found	Set `planReview.gaps` , auto-loop to plan
Plan-review approved	Set `planReview.approved = true` , phase = "delegate"
Task dispatched	Set task `status = "in_progress"` , `startedAt`
Task complete	Set task `status = "complete"` , `completedAt`
Worktree created	Add to `worktrees` object
Review complete	Update `reviews` object
PR created	Set `artifacts.pr` , `synthesis.prUrl`
PR feedback	Append to `synthesis.prFeedback`

事件	状态更新操作
`/ideate` 启动	创建状态文件
设计文档保存	设置 `artifacts.design` ，阶段设为"plan"
计划文档保存	设置 `artifacts.plan` ，填充tasks，阶段设为"plan-review"
计划评审发现问题	设置 `planReview.gaps` ，自动循环回plan阶段
计划评审通过	设置 `planReview.approved = true` ，阶段设为"delegate"
任务分派	设置任务 `status = "in_progress"` ，记录 `startedAt`
任务完成	设置任务 `status = "complete"` ，记录 `completedAt`
工作树创建	添加至 `worktrees` 对象
评审完成	更新 `reviews` 对象
PR创建	设置 `artifacts.pr` 、 `synthesis.prUrl`
PR反馈	追加至 `synthesis.prFeedback`

Oneshot-specific state updates

Oneshot专属状态更新

Oneshot is a first-class workflow type with a compressed lifecycle and an opt-in PR path. The rows below mirror the feature-workflow table above.

Phase	State updates	Events emitted
`plan` (oneshot)	`oneshot.planSummary` , `artifacts.plan` , optional `oneshot.synthesisPolicy`	`workflow.transition`
`implementing` (oneshot)	`tasks[].status` , `artifacts.tests`	`task.*` , optional `synthesize.requested` (via `request_synthesize` )
`synthesize` (oneshot)	`synthesis.prUrl` , `artifacts.pr`	`workflow.transition` , `stack.submitted`
`completed` (oneshot)	—	`workflow.transition` (to `completed` )

The

implementing → synthesize | completed

fork is a choice state resolved by

finalize_oneshot

, which reads the

synthesisOptedIn

guard (

synthesisPolicy

synthesize.requested

events). See

@skills/oneshot-workflow/SKILL.md

for the full opt-in mechanics.

Oneshot是一等工作流类型，具有压缩的生命周期和可选PR路径。以下表格与上方功能工作流表格对应。

阶段	状态更新	触发事件
`plan` （oneshot）	`oneshot.planSummary` 、 `artifacts.plan` 、可选 `oneshot.synthesisPolicy`	`workflow.transition`
`implementing` （oneshot）	`tasks[].status` 、 `artifacts.tests`	`task.*` 、可选 `synthesize.requested` （通过 `request_synthesize` 触发）
`synthesize` （oneshot）	`synthesis.prUrl` 、 `artifacts.pr`	`workflow.transition` 、 `stack.submitted`
`completed` （oneshot）	—	`workflow.transition` （切换至 `completed` ）

implementing → synthesize | completed

分支是由

finalize_oneshot

解析的选择状态，该函数会读取

synthesisOptedIn

防护条件（

synthesisPolicy

synthesize.requested

事件）。关于完整的可选PR机制，请参考

@skills/oneshot-workflow/SKILL.md

。

Automatic State Updates

自动状态更新

Skills should update state at key moments:

brainstorming/SKILL.md:

markdown

After saving design:
1. Update state: `.artifacts.design = "<path>"`
2. Update state: `.phase = "plan"`

implementation-planning/SKILL.md:

markdown

After saving plan:
1. Update state: `.artifacts.plan = "<path>"`
2. Populate `.tasks` from plan
3. Update state: `.phase = "delegate"`

delegation/SKILL.md:

markdown

On task dispatch:
- Update task status to "in_progress"
- Add worktree to state if created

On task complete:
- Update task status to "complete"
- Check if all tasks done, suggest checkpoint

Skills应在关键节点更新状态：

brainstorming/SKILL.md：

markdown

保存设计文档后：
1. 更新状态：`.artifacts.design = "<path>"`
2. 更新状态：`.phase = "plan"`

implementation-planning/SKILL.md：

markdown

保存计划文档后：
1. 更新状态：`.artifacts.plan = "<path>"`
2. 从计划中填充`.tasks`
3. 更新状态：`.phase = "delegate"`

delegation/SKILL.md：

markdown

任务分派时：
- 将任务状态更新为"in_progress"
- 若创建了工作树则添加至状态

任务完成时：
- 将任务状态更新为"complete"
- 检查是否所有任务完成，建议创建检查点

State Schema

状态模式

See

docs/schemas/workflow-state.schema.json

for full schema.

Key sections:

```
version
```
: Schema version (currently "1.1")
```
featureId
```
: Unique workflow identifier
```
workflowType
```
: Required. One of "feature", "debug", "refactor", or "oneshot"
```
phase
```
: Current workflow phase
```
artifacts
```
: Paths to design, plan, PR
```
tasks
```
: Task list with status
```
worktrees
```
: Active git worktrees
```
planReview
```
: Plan-review delta analysis results (
```
gaps
```
,
```
approved
```
)
```
reviews
```
: Review results
```
synthesis
```
: Merge/PR state

完整模式请参考

docs/schemas/workflow-state.schema.json

。

关键部分：

```
version
```
：模式版本（当前为"1.1"）
```
featureId
```
：唯一工作流标识符
```
workflowType
```
：必填项。可选值为"feature"、"debug"、"refactor"或"oneshot"
```
phase
```
：当前工作流阶段
```
artifacts
```
：设计文档、计划文档、PR的路径
```
tasks
```
：带状态的任务列表
```
worktrees
```
：活跃的git工作树
```
planReview
```
：计划评审差异分析结果（
```
gaps
```
、
```
approved
```
）
```
reviews
```
：评审结果
```
synthesis
```
：合并/PR状态

Best Practices

最佳实践

Update often - State should reflect reality at all times
Use MCP tools - Prefer workflow-state MCP tools over manual JSON editing
Reconcile on resume - Always verify state matches git state
Checkpoint at boundaries - Save state before likely context exhaustion
Read state, don't remember - After summarization, read from state file

频繁更新 - 状态应始终反映实际情况
使用MCP工具 - 优先使用工作流状态MCP工具，而非手动编辑JSON
恢复时对账 - 始终验证状态与git状态一致
边界处创建检查点 - 在可能出现上下文耗尽前保存状态
读取状态，而非记忆 - 总结后从状态文件读取信息

Troubleshooting

故障排查

MCP Tool Call Failed

MCP工具调用失败

If an Exarchos MCP tool returns an error:

Check the error message — it usually contains specific guidance
Verify the workflow state exists: call
```
mcp__exarchos__exarchos_workflow
```
with
```
action: "get"
```
and the featureId
If "version mismatch": another process updated state — retry the operation

If state is corrupted: call

mcp__exarchos__exarchos_workflow

with

action: "cancel"

and

dryRun: true

如果Exarchos MCP工具返回错误：

查看错误信息——通常包含具体指导
验证工作流状态是否存在：调用
```
mcp__exarchos__exarchos_workflow
```
并设置
```
action: "get"
```
和对应的featureId
若提示"version mismatch"：其他进程已更新状态——重试操作

若状态损坏：调用

mcp__exarchos__exarchos_workflow

并设置

action: "cancel"

和

dryRun: true

State Desync

状态不同步

If workflow state doesn't match git reality:

The SessionStart hook runs reconciliation automatically on resume
If manual check needed: compare state file with
```
git log
```
and branch state

Update state via

mcp__exarchos__exarchos_workflow

with

action: "set"

to match git truth

如果工作流状态与git实际情况不符：

SessionStart钩子会在恢复时自动执行对账
如需手动检查：对比状态文件与
```
git log
```
及分支状态
使用
```
mcp__exarchos__exarchos_workflow
```
并设置
```
action: "set"
```
更新状态，使其与git实际情况一致

Checkpoint File Missing

检查点文件缺失

If the PreCompact hook can't find state to checkpoint:

Verify a workflow is active: call
```
mcp__exarchos__exarchos_workflow
```
with
```
action: "get"
```
and the featureId
If no active workflow: the hook will silently skip (expected behavior)
If workflow exists but checkpoint fails: check disk space and permissions

如果PreCompact钩子无法找到用于创建检查点的状态：

验证是否存在活跃工作流：调用
```
mcp__exarchos__exarchos_workflow
```
并设置
```
action: "get"
```
和对应的featureId
若无活跃工作流：钩子会静默跳过（预期行为）
若工作流存在但检查点创建失败：检查磁盘空间和权限

Resume Finds Stale State

恢复时发现过期状态

If state references branches or worktrees that no longer exist:

The SessionStart hook handles reconciliation automatically
It updates state to reflect current git reality
Missing branches are flagged in the session-start output

如果状态引用的分支或工作树已不存在：

SessionStart钩子会自动处理对账
它会更新状态以反映当前git实际情况
缺失的分支会在会话启动输出中标记

Multiple Active Workflows

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

mcp__exarchos__exarchos_workflow

with

action: "cancel"

and

dryRun: true

on stale workflows to preview cleanup

Cancel stale workflows before starting new ones

如果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

对过期工作流调用

mcp__exarchos__exarchos_workflow

并设置

action: "cancel"

和

dryRun: true

，预览清理操作

在启动新工作流前取消过期工作流

Example Workflow

示例工作流

Start new workflow: Use

mcp__exarchos__exarchos_workflow

with

action: "init"

with

featureId: "user-authentication"

workflowType: "feature"

After design phase: Use

mcp__exarchos__exarchos_workflow

with

action: "set"

with:

```
featureId: "user-authentication"
```
```
phase: "plan"
```

updates: { "artifacts.design": "docs/designs/2026-01-05-user-auth.md" }

Check state: Use

mcp__exarchos__exarchos_workflow

with

action: "get"

with

featureId: "user-authentication"

Resume after context loss: Use

mcp__exarchos__exarchos_workflow

with

action: "get"

with

featureId: "user-authentication"

to get context restoration output

启动新工作流：使用

mcp__exarchos__exarchos_workflow

并设置

action: "init"

，同时传入

featureId: "user-authentication"

、

workflowType: "feature"

设计阶段完成后：使用

mcp__exarchos__exarchos_workflow

并设置

action: "set"

，同时传入：

```
featureId: "user-authentication"
```
```
phase: "plan"
```

updates: { "artifacts.design": "docs/designs/2026-01-05-user-auth.md" }

检查状态：使用

mcp__exarchos__exarchos_workflow

并设置

action: "get"

和

featureId: "user-authentication"

上下文丢失后恢复：使用

mcp__exarchos__exarchos_workflow

并设置

action: "get"

和

featureId: "user-authentication"

获取上下文恢复输出

workflow-state

Original

Translation

Workflow State Management Skill

工作流状态管理Skill

Overview

概述

Triggers

触发条件

Phase Transitions

阶段切换

Schema Discovery

模式发现

State Location

状态存储位置

State Operations

状态操作

Initialize State

初始化状态

Workflow Types at a Glance

工作流类型概览

Read State

读取状态

Update State

更新状态

Editing the tasks array

编辑tasks数组

Get Summary

获取摘要

Reconcile State

状态对账

Integration Points

集成点

When to Update State

状态更新时机

Oneshot-specific state updates

Oneshot专属状态更新

Automatic State Updates

自动状态更新

State Schema

状态模式

Best Practices

最佳实践

Troubleshooting

故障排查

MCP Tool Call Failed

MCP工具调用失败

State Desync

状态不同步

Checkpoint File Missing

检查点文件缺失

Resume Finds Stale State

恢复时发现过期状态

Multiple Active Workflows

多个活跃工作流

Example Workflow

示例工作流

Editing the
`tasks`
array

编辑
`tasks`
数组