Back to Details

session-checkpoint

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Session Checkpoint

会话检查点

Commands

命令

  • Save: 
    save checkpoint {name}
    , 
    save checkpoint
    (prompts for descriptive name)
  • Resume: 
    continue
    , 
    continue {name}
    , 
    resume
    (auto-select if one checkpoint, list if many)
  • Cleanup: 
    clear checkpoint {name}
    , 
    clear all checkpoints
  • 保存： 
    save checkpoint {name}
    、
    save checkpoint
    （会提示输入描述性名称）
  • 恢复： 
    continue
    、
    continue {name}
    、
    resume
    （如果只有一个检查点则自动选择，多个则列出选项）
  • 清理： 
    clear checkpoint {name}
    、
    clear all checkpoints

Save Procedure

保存流程

1. Determine Name

1. 确定名称

If user gave a name, use it (lowercase, sanitize: replace 
/
with 
-
).
If no name given:
  • Pick a short name from the current work (2-4 words, e.g., "my-feature")
  • Confirm with 
    AskUserQuestion
    , options:
    • "Use '{suggested-name}' (Recommended)"
    • "Provide different name" (user types via Other)
Sanitize final name: lowercase, replace spaces/
/
with 
-
, strip leading 
-
.
如果用户提供了名称，直接使用（转为小写，清理格式：将
/
替换为
-
）。
如果未提供名称：
  • 从当前工作内容中选取一个短名称（2-4个词，例如"my-feature"）
  • 通过
    AskUserQuestion
    确认，选项如下：
    • "使用'{suggested-name}'（推荐）"
    • "提供其他名称"（用户通过“其他”输入）
最终名称清理规则：转为小写，将空格和
/
替换为
-
，去除开头的
-
。

2. Locate Memory Directory

2. 定位内存目录

Find the auto memory directory path from the system prompt line: 
You have a persistent auto memory directory at <path>
从系统提示行中找到自动内存目录路径： 
You have a persistent auto memory directory at <path>

3. Gather State

3. 收集状态

Run in a single bash call:
bash
git branch --show-current && git status --short | head -10 && date '+%Y-%m-%d %H:%M'
From conversation context, determine:
  • What was accomplished this session
  • What failed or was abandoned (and why)
  • Where work left off — don't infer next steps, but DO capture all actionable context the next session needs: pending decisions, proposed options, open questions
  • Which files were modified
在一次bash调用中执行以下命令：
bash
git branch --show-current && git status --short | head -10 && date '+%Y-%m-%d %H:%M'
从对话上下文确定：
  • 本次会话已完成的内容
  • 失败或放弃的尝试（以及原因）
  • 工作中断的位置——不要推断下一步，但必须记录下一次会话所需的所有可操作上下文：待决策事项、提议的选项、未解决的问题
  • 哪些文件被修改

4. Find Active Plan

4. 查找活跃计划

A plan file exists whenever the session involved plan mode or "implement this plan." Look for it in this order:
  1. System prompt: When plan mode is active, the system prompt contains the plan file path (e.g., 
    ~/.claude/plans/{name}.md
    ). Check for it.
  2. Conversation context: The user's message may reference or paste a plan, or mention "implement the following plan." The plan file path may also appear in an 
    ExitPlanMode
    result or plan-related system message.
  3. Fallback: If the session clearly involved executing a multi-step plan but no file path was found, check 
    ~/.claude/plans/
    for the most recently modified 
    .md
    file, read its first lines, and confirm it matches the work being done.
If a plan is found, read it to determine step N of M (count total tasks/steps, identify which one work stopped at).
Default: include the Plan line. Only omit it if the session genuinely had no plan — a simple task, a one-off question, etc. Err on the side of including it.
当会话涉及计划模式或"implement this plan"时，会存在一个计划文件。按以下顺序查找：
  1. 系统提示： 当计划模式激活时，系统提示中会包含计划文件路径（例如
    ~/.claude/plans/{name}.md
    ），请检查该路径。
  2. 对话上下文： 用户的消息可能引用或粘贴了计划，或者提到"implement the following plan"。计划文件路径也可能出现在
    ExitPlanMode
    结果或与计划相关的系统消息中。
  3. 备选方案： 如果会话明显涉及执行多步骤计划但未找到文件路径，请检查
    ~/.claude/plans/
    目录中最近修改的
    .md
    文件，读取其前几行并确认是否与当前工作匹配。
如果找到计划文件，读取它以确定当前是第N步（共M步）：统计总任务/步骤数，确定工作中断在哪个步骤。
默认规则：包含计划行。 只有当会话确实没有计划时（例如简单任务、一次性问题等）才省略该行。尽量优先包含计划行。

5. Write Checkpoint File

5. 写入检查点文件

If 
checkpoint-{name}.md
already exists, use 
AskUserQuestion
to confirm — show the existing checkpoint's Saved date and Left Off summary, with options: "Overwrite" / "Different name" (user provides new name via "Other"). If they choose a different name, continue from this step with the new name.
Write 
<memory_dir>/checkpoint-{name}.md
:
markdown
undefined
如果
checkpoint-{name}.md
已存在，使用
AskUserQuestion
确认——显示现有检查点的保存日期和中断位置摘要，提供选项：“覆盖” / “使用其他名称”（用户通过“其他”输入新名称）。如果用户选择其他名称，则使用新名称重新执行此步骤。
写入文件
<memory_dir>/checkpoint-{name}.md
：
markdown
undefined

Checkpoint: {name}

Checkpoint: {name}

  • Branch: 
    {branch}
  • Saved: {YYYY-MM-DD HH:MM}
  • Plan: 
    ~/.claude/plans/{file}
    (step N of M)
  • Branch: 
    {branch}
  • Saved: {YYYY-MM-DD HH:MM}
  • Plan: 
    ~/.claude/plans/{file}
    (step N of M)

Left Off

Left Off

{1-3 lines describing where work stopped}
{1-3 lines describing where work stopped}

Done This Session

Done This Session

  • {accomplishments}
  • {accomplishments}

Failed Approaches

Failed Approaches

  • {what + why — or "None"}
  • {what + why — or "None"}

Modified Files

Modified Files

  • {list}

Omit the **Plan:** line if no plan file exists.
  • {list}

如果没有计划文件，则省略**Plan:**行。

6. Update MEMORY.md Index

6. 更新MEMORY.md索引

Read MEMORY.md. Find the 
## Active Checkpoints
section.
  • If checkpoint name already in index: update its line.
  • If not: add a new bullet.
  • If no 
    ## Active Checkpoints
    section exists: insert it after the first 
    #
    heading. If MEMORY.md doesn't exist, create it with 
    # Project Memory
    as the heading.
Bullet format:
- **{name}** ({branch}, {Mon DD}) — {one-line summary}
Keep the 
Resume any: `continue` or `continue {name}`
line after the bullets.
Do NOT touch any other sections in MEMORY.md.
读取MEMORY.md文件，找到
## Active Checkpoints
部分：
  • 如果检查点名称已在索引中：更新对应的行。
  • 如果不存在：添加一个新的项目符号。
  • 如果没有
    ## Active Checkpoints
    部分：在第一个
    #
    标题后插入该部分。如果MEMORY.md不存在，则创建该文件并将
    # Project Memory
    作为标题。
项目符号格式：
- **{name}** ({branch}, {Mon DD}) — {one-line summary}
在项目符号后保留
Resume any: `continue` or `continue {name}`
这一行。
不要修改MEMORY.md中的其他部分。

7. Output

7. 输出内容

Keep output minimal — this often runs near end of context window.
Checkpoint "{name}" saved. Resume anytime: `continue {name}`
That's it. Don't ask follow-up questions. Don't suggest 
/compact
or 
/clear
— the user manages their own context window.
输出内容应尽量简洁——因为此操作通常在上下文窗口接近满时执行。
Checkpoint "{name}" saved. Resume anytime: `continue {name}`
操作完成。不要提出后续问题，不要建议使用
/compact
或
/clear
——用户会自行管理其上下文窗口。

Resume Procedure

恢复流程

1. Determine Name

1. 确定名称

If user gave a name, use it. If no name given:
  • List 
    checkpoint-*.md
    files in memory directory
  • If exactly one: use it automatically
  • If multiple: use 
    AskUserQuestion
    with checkpoint names as options (include branch and date in each option's description)
  • If none: report "No saved checkpoints found." and stop
如果用户提供了名称，直接使用。 如果未提供名称：
  • 列出内存目录中的
    checkpoint-*.md
    文件
  • 如果只有一个：自动选择该文件
  • 如果有多个：使用
    AskUserQuestion
    ，将检查点名称作为选项（每个选项的描述中包含分支和日期）
  • 如果没有：提示“No saved checkpoints found.”并停止操作

2. Read Checkpoint File

2. 读取检查点文件

Read 
<memory_dir>/checkpoint-{name}.md
. If file missing: remove the orphaned bullet from MEMORY.md's 
## Active Checkpoints
, report "Checkpoint file missing (cleaned up stale entry)", and stop.
读取
<memory_dir>/checkpoint-{name}.md
文件。 如果文件不存在：从MEMORY.md的
## Active Checkpoints
中移除对应的孤立项目符号，提示“Checkpoint file missing (cleaned up stale entry)”并停止操作。

3. Check Branch

3. 检查分支

bash
git branch --show-current
If current branch differs from checkpoint's branch, warn:
Warning: You're on `{current}` but checkpoint was saved on `{saved}`. Switch with: git checkout {saved}
Continue regardless — user decides.
bash
git branch --show-current
如果当前分支与检查点的分支不同，发出警告：
Warning: You're on `{current}` but checkpoint was saved on `{saved}`. Switch with: git checkout {saved}
无论如何继续操作——由用户决定是否切换分支。

4. Load Plan

4. 加载计划

If checkpoint references a plan file, read it.
如果检查点引用了计划文件，则读取该文件。

5. Present and Begin

5. 展示内容并开始

Output the checkpoint summary and last context. If Failed Approaches is non-empty, include them:
⚠ Previously failed: {approach} — {why}
输出检查点摘要和最后的上下文。如果Failed Approaches非空，则包含以下内容：
⚠ Previously failed: {approach} — {why}

6. Auto-Clear

6. 自动清理

The checkpoint is now consumed — context is live. Delete the checkpoint file and remove its bullet from MEMORY.md's 
## Active Checkpoints
(same as Cleanup steps 1-2). If no bullets remain, remove the entire section.
Do NOT mention the cleanup to the user. They see only the summary from step 5, then wait for their direction.
检查点已被使用——上下文已激活。删除检查点文件，并从MEMORY.md的
## Active Checkpoints
中移除对应的项目符号（与清理流程的步骤1-2相同）。如果没有剩余的项目符号，则移除整个部分。 不要向用户提及清理操作。用户只会看到步骤5中的摘要，然后等待他们的指示。

Cleanup Procedure

清理流程

1. Delete File

1. 删除文件

Remove the checkpoint file via Bash:
bash
rm <memory_dir>/checkpoint-{name}.md
通过Bash删除检查点文件：
bash
rm <memory_dir>/checkpoint-{name}.md

2. Update MEMORY.md

2. 更新MEMORY.md

Remove the bullet for 
{name}
from 
## Active Checkpoints
. If no bullets remain, remove the entire section (header + resume line).
从
## Active Checkpoints
中移除
{name}
对应的项目符号。 如果没有剩余的项目符号，则移除整个部分（标题+恢复提示行）。

3. Output

3. 输出内容

Cleared checkpoint "{name}"
Cleared checkpoint "{name}"

clear all checkpoints

clear all checkpoints
命令

Glob 
checkpoint-*.md
in memory directory. Delete all matches. Remove the entire 
## Active Checkpoints
section (header + bullets + resume line) from MEMORY.md. Output:
Cleared {N} checkpoint(s)
匹配内存目录中的
checkpoint-*.md
文件，删除所有匹配项。从MEMORY.md中移除整个
## Active Checkpoints
部分（标题+项目符号+恢复提示行）。输出：
Cleared {N} checkpoint(s)