workbuddy迁移缓存到自定义位置

SKILL运维和基础设施
运维和基础设施 未标注权限 许可证不明确
安装这个技能

修改workbuddy缓存目录,释放C盘空间

一句话选型结论

按资料字段整理

workbuddy迁移缓存到自定义位置是运维和基础设施类的 Agent Skill,安装方式未标注,权限级别未标注,许可证不明确,有可直接复制的安装命令。技能本质是给模型读的说明书,效果取决于模型与客户端;装完先用一个小任务验证,再放进正式流程。

分类运维和基础设施 安装方式未标注 权限级别未标注权限 许可证许可证不明确
累计安装 未标注 本周 未标注
GitHub 星标 未标注 仓库 未标注
下载量 未标注 技能市场下载数
权限级别 未标注权限 安装前请确认风险范围
许可证 许可证不明确 未说明

安装

未标注
在支持 Skills 的客户端里执行,装完问一句「你有哪些技能」验证
skillhub install user_de70d336/zb-workbuddyskill01
  1. 1
    确认客户端支持

    确认你的客户端支持 Skills(Claude、Codex、Cursor 等),并升级到较新的版本。

  2. 2
    拿到技能文件

    执行上面的命令安装;也可以把仓库里的 SKILL.md 手动放进客户端的 skills 目录。

  3. 3
    刷新技能列表

    重新打开客户端,或刷新技能列表,让新放进来的技能被加载。

  4. 4
    验证是否生效

    问一句「你有哪些技能」,确认列表里出现它,再拿一个小任务试跑。

技能档案

运维和基础设施
字段来自技能仓库与公开索引,中文说明由本站标注
作者 / 维护方user_de70d336仓库 未标注 · 版本 未标注
宿主客户端未标注技能可被这些客户端加载
安装方式未标注安装包 community
权限级别未标注权限安装前请确认风险范围
许可证未说明许可证不明确,商用前请自行确认条款
仓库可用性available技能仓库当前是否可访问
最近核验未标注上游最后核验时间 · 数据更新 1791671926
正文体量8.3 KBSKILL.md 原文大小,越大说明说明书写得越细

典型使用场景

3 类
按分类整理的常见用法,具体以 SKILL.md 为准
查服务与日志

按现象定位服务状态与日志线索。

写部署与巡检脚本

把重复的部署、巡检写成脚本。

故障排查

按步骤排查网络、依赖与配置。

选择建议

装之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有可直接复制的安装命令,不用自己翻文档找入口。

先想清楚的地方

  • 没有标注权限级别,实际能做什么要装好后确认。
  • 许可证不明确,商用或二次分发前先到仓库确认条款。
  • 没有做站内核验,内容以作者仓库为准。
  • 技能只是说明书,实际效果取决于模型与客户端,装完先用小任务试一遍。

更适合

  • 需要在运维和基础设施场景里给模型补一套做法的人。
  • 愿意先读一遍 SKILL.md、确认它会让模型做什么的人。

不太适合

  • 对许可证要求很严、必须明确可商用的人。
  • 只想装完就自动变强、不打算看说明书的人。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
技能权限级别安装方式累计安装宿主客户端安装命令
workbuddy迁移缓存到自定义位置当前 未标注权限 未标注 未标注 未标注 有
Azure 部署 调用外部服务 GitHub 仓库 664.2 万 Codex、Claude、Cursor 有
Azure diagnostics 问题管理 调用外部服务 GitHub 仓库 657.6 万 Codex、Claude、Cursor 有
Azure aigateway 部署 调用外部服务 GitHub 仓库 655.3 万 Codex、Claude、Cursor 有
Azure AI 搜索 调用外部服务 GitHub 仓库 651.4 万 Codex、Claude、Cursor 有
Azure Cloud migrate 部署 调用外部服务 GitHub 仓库 650.7 万 Codex、Claude、Cursor 有

SKILL.md 全文

8.3 KB
模型实际读到的说明书原文,来自技能仓库

WorkBuddy 缓存位置迁移 / WorkBuddy Cache Relocate

Overview

Move the WorkBuddy data root (default %USERPROFILE%\.workbuddy on Windows, ~/.workbuddy on macOS/Linux)
to another drive or folder, and keep skills, plugins and settings loading correctly afterwards.

The non-obvious part is why this is not simply an environment variable change. WorkBuddy resolves its
root from WORKBUDDY_CONFIG_DIR / CODEBUDDY_CONFIG_DIR, so setting those variables does redirect
new writes. But the app also persists absolute paths into several live state files - every plugin's
installPath, marketplace locations, skill metadata, the skill picker index, and the threat database.
After a naive variable change the app reads its manifests from the new location while those manifests
still point at the old one, which silently loads stale copies or fails outright.

The fix is to keep the old path valid by turning it into a directory junction (Windows) or
symlink (macOS/Linux) that points at the new location. Both mechanisms then agree, every historical
absolute path keeps resolving, and only one physical copy of the data exists.

When to use

  • The system drive is nearly full and WorkBuddy's data root is a major consumer
  • The user wants skills, plugins, logs or config stored on another drive
  • The user asks how to change CODEBUDDY_CONFIG_DIR / WORKBUDDY_CONFIG_DIR
  • The user asks to move or migrate .workbuddy

Preconditions

  • Sufficient free space on the target drive, confirmed in Step 1
  • Ability to create user-level environment variables
  • Steps 1 and 2 run safely while WorkBuddy is open
  • Step 3 requires WorkBuddy to be fully closed - including the tray icon. Never attempt it live.

Workflow

Run the steps in order. Do not skip Step 1: it decides whether the junction is required and surfaces
blockers before anything is changed.

Step 1 - Diagnose (safe, runs live)

python <skill-dir>/scripts/diagnose.py

Reports the persisted and in-process environment variables, the resolved root, whether that root is
already a junction, per-directory sizes, free space on both drives, and every live state file holding
an absolute path into the old root.

Stop and report instead of continuing if the root is already a junction - the migration has been done.
Stop if free space on the target is less than the reported root size plus 20%.

Step 2 - Prepare: full copy and environment variables (safe, runs live)

python <skill-dir>/scripts/prepare_migration.py --target D:\workbuddydata

Perform a full copy, then set WORKBUDDY_CONFIG_DIR and CODEBUDDY_CONFIG_DIR at user scope to the
target. Files locked by the running app are expected to be skipped here; Step 3 syncs them.

This step is non-destructive: it writes to the target and to the registry (or shell profile), and
never modifies or deletes the original directory. Report what was copied and which files were skipped.

Step 3 - Finalize: delta sync and junction (WorkBuddy MUST be closed)

Hand this step to the user rather than executing it.

Two things must happen before the script can run, and both are easy to get wrong:

  1. The script ships as plain text. It is scripts/finalize_migration.cmd.txt because SkillHub's
    upload whitelist does not include .cmd. Copy it to a location the user controls and save it as
    finalize_migration.cmd - the user's Desktop is a good choice. Windows will ask whether to change
    the extension; that prompt must be confirmed.
  2. That location must be outside the data root. The script renames the data root, so it cannot
    live inside it. That rules out running it straight from the skill folder at
    <root>\skills\workbuddy-cache-relocate\scripts\. The script checks this itself and refuses with a
    clear message, but copying it out first is the intended flow.

Then instruct the user to quit WorkBuddy completely - tray icon included - confirm that no
WorkBuddy.exe remains in Task Manager, and double-click the script (or run it from a terminal):

finalize_migration.cmd <target>
finalize_migration.cmd <target> <source>      (optional 2nd arg)

No Python is needed for this step - it is pure cmd. That matters: the interpreter bundled with
WorkBuddy is not on the persisted PATH, so python does not resolve once the app is closed. Never
tell the user to run a .py at this point.

The source defaults to %USERPROFILE%\.workbuddy. Pass the second argument only when the data lives
somewhere else, which is the state left behind by setting the environment variables by hand without a
junction.

The script:

  • refuses to run if it detects that it is sitting inside the data root
  • self-tests the process-detection mechanism first, and refuses to continue if it cannot read the
    process list, rather than risk a rename while the app is still writing
  • refuses to run while WorkBuddy is alive
  • verifies that the source is a real directory and that the target already holds settings.json
  • performs the delta sync
  • renames the original folder to <source>.old
  • creates the junction and verifies that settings.json and skills are reachable through it

Any failure after the rename is rolled back automatically, leaving the original folder in place. It is
safe to re-run: an existing junction is detected and reported rather than acted on.

On macOS or Linux use the symlink procedure in references/internals.md instead.

Step 4 - Verify

After the user restarts WorkBuddy, confirm all of the following:

  • The skill list and plugin list load normally
  • C:\Users\<user>\.workbuddy\skills resolves to the target drive
  • No new files appear under the old path other than through the junction
  • The app still reports the same account and settings

If plugins or skills fail to load, check references/internals.md before rolling back.

Step 5 - Cleanup

The old data is retained as .workbuddy.old next to the original path. Advise the user to keep it
until WorkBuddy has run without problems for several days, then delete it to reclaim the space. Deleting
it is the step that actually frees the system drive. Never delete it automatically.

Rollback

  • Before Step 3: delete the target directory. The original is untouched.
  • After Step 3: remove the junction with rmdir "C:\Users\<user>\.workbuddy" (no /s, so only the link
    is removed), rename .workbuddy.old back, and clear the two environment variables.
  • The finalize script performs the rollback automatically when a step fails.

Resources

  • scripts/diagnose.py - read-only inspection of root, sizes, free space and stale absolute paths
  • scripts/prepare_migration.py - full copy plus user-level environment variables
  • scripts/finalize_migration.cmd.txt - delta sync plus junction, with automatic rollback. Shipped as
    .txt because .cmd is outside the SkillHub upload whitelist; copy it outside the data root,
    rename it to .cmd and run it with WorkBuddy closed
  • references/internals.md - how the root is resolved, the state files that hold absolute paths, the
    macOS/Linux symlink procedure, sizing reference and troubleshooting

常见问题

5 问
答案由本站依据条目字段整理
workbuddy迁移缓存到自定义位置怎么安装?

在支持 Skills 的客户端里执行:skillhub install user_de70d336/zb-workbuddyskill01。安装方式为未标注。装完刷新技能列表或重开客户端。

workbuddy迁移缓存到自定义位置支持哪些客户端?

资料没有标注宿主客户端,但只要客户端支持 Skills(Claude、Codex、Cursor 等),一般都能加载。

workbuddy迁移缓存到自定义位置权限高吗,安全吗?

权限级别为未标注权限。安装前先看它会读取或写入哪些内容。

workbuddy迁移缓存到自定义位置收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于技能本身和所用的模型或客户端,正式使用前按作者说明确认。

workbuddy迁移缓存到自定义位置用什么许可证?

许可证标注为 未说明(许可证不明确)。商用或二次分发前请到仓库确认条款原文。

热门专题

按场景挑技能
全部专题 →

技能内容以作者仓库为准;分类、权限级别与许可证状态的中文说明为中文整理,最近核验 未标注。