Skip to main content
Agent skill

Skillfs Mount

Help users configure, mount, and unmount a skillfs FUSE virtual filesystem on their local machine (e.g. for openclaw). Covers analyzing existing skill usage to suggest view configuration, generating skillfs-views.toml, mounting, and graceful unmount. Use this skill when the user asks to: set up skillfs, configure skill views, mount or unmount skillfs, check if skillfs is running, or get help organizing their skills directory (especially for openclaw users).

Get started

Get the complete skill folder, including its scripts and reference files.

When to use this skill

Use Skillfs Mount when an AI agent needs a reusable SKILL.md workflow for this job: Help users configure, mount, and unmount a skillfs FUSE virtual filesystem on their local machine (e.g. for openclaw). Covers analyzing existing skill usage to suggest view configuration, generating skillfs-views.toml, mounting, and graceful unmount. Use this skill when the user asks to: set up skillfs, configure skill views, mount or unmount skillfs, check if skillfs is running, or get help organizing their skills directory (especially for openclaw users).

When not to use it

Skip Skillfs Mount when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

Where will you use it?

Install scope
Claude Code folder setup
mkdir -p ~/.claude/skills/skillfs-mount
# Save SKILL.md to ~/.claude/skills/skillfs-mount/SKILL.md
  1. Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/skillfs-mount/SKILL.md. For a shared project, commit the folder at .claude/skills/skillfs-mount/ instead.
  2. Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
  3. For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.

Full install guide for Claude, Cursor, and Codex →

Try a starter prompt

After installing the skill, paste this starter prompt into your agent to verify it triggers on a small task.

Starter prompt
Use the Skillfs Mount skill for this task: Use Skillfs Mount when this documented workflow matches the task. Confirm any required inputs or files before making changes.

What supports this recommendation?

71/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.

Documentation
77/100
Practical detail
68/100
Source evidence
63/100
Source signals
72/100
Model compatibility evidence
  • claudeuntested

    No model-specific signal or recorded compatibility test was found.

  • chatgptuntested

    No model-specific signal or recorded compatibility test was found.

  • geminiuntested

    No model-specific signal or recorded compatibility test was found.

  • copilotinferred

    The skill text mentions Copilot or a closely associated term. This is keyword-derived compatibility; no hands-on model test is recorded.

  • llamauntested

    No model-specific signal or recorded compatibility test was found.

  • perplexityuntested

    No model-specific signal or recorded compatibility test was found.

  • mistraluntested

    No model-specific signal or recorded compatibility test was found.

  • grokuntested

    No model-specific signal or recorded compatibility test was found.

Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.

What this skill does

# skillfs 配置与挂载指南

## 前提条件

- `skillfs` 已安装(RPM 包),运行 `skillfs --version` 确认 - 系统已安装 `fuse3`(若缺失:`yum install -y fuse3`) - 确认 skills 源目录路径(openclaw 用户通常为 `~/.openclaw/skills`)

---

## 操作流程

### STEP 1:了解用户类型,确定主视图

**有使用历史的用户(openclaw):**

运行脚本分析历史调用频次,按调用次数决定哪些 skill 进主视图:

```bash # 分析 openclaw session 日志中的 skill 使用频次(默认读 ~/.openclaw) python3 <skill_dir>/scripts/skill_usage_from_session_logs.py

# 分析 copilot-shell (cosh) chat 日志中的 skill 调用统计(默认读 ~/.copilot-shell) python3 <skill_dir>/scripts/skill_usage_from_chat_logs.py ```

两个脚本都支持 `--logs-dir` 指向自定义日志目录。

输出示例: ``` === Skill Invocation Statistics === github: 42 slack: 28 tmux: 17 weather: 3 ```

调用频次高的 skill 建议放入主视图,频次低或从未使用的放入次级视图。

**纯新用户(无历史):**

列出 skills 目录内容,按用途给出分类建议:

```bash ls ~/.openclaw/skills/ ```

常见分类建议: - 主视图(高频核心):代码工具、版本管理、常用 API 集成 - 次级视图:不常用工具、实验性 skill、特定场景用途

---

### STEP 2:生成 skillfs-views.toml

根据分析结果,运行 classify 命令生成初始配置:

```bash skillfs classify ~/.openclaw/skills --primary-count 8 ```

生成的 `~/.openclaw/skills/skillfs-views.toml` 结构(视图名称固定为 `major` / `other`):

```toml [[view]] name = "major" default = true description = "Core skills shown at mount time" skills = ["github", "slack", "tmux", ...]

[[view]] name = "other" default = false description = "Additional skills accessible via skill-discover" skills = ["weather", "notion", ...] ```

> 挂载行为由 `default = true/false` 决定,视图名称仅作为 skill-discover 中的分组标题显示。

**按分析结果手动调整**:直接编辑 views.toml,在两个 skills 列表之间移动 skill 名称。

> ⚠️ views.toml 中的字符串必须与 skill 的**目录名**完全一致,否则该技能将从视图中 > 消失。目录名是权威的 skill key;`SKILL.md` frontmatter 里的 `name:` 只是展示元数据, > 不参与视图匹配,重命名目录后也不会覆盖目录 key。

---

### STEP 3:挂载

```bash # 原地挂载(推荐):Agent 访问同一路径,FUSE 透明过滤内容 skillfs mount ~/.openclaw/skills ~/.openclaw/skills \ --pid-file /tmp/skillfs.pid \ --log-file /tmp/skillfs-{pid}.log & sleep 1

# 确认挂载成功 ls ~/.openclaw/skills/ # 应只显示主视图 skill + skill-discover ```

挂载时将未分配到任何视图的新 skill 纳入内存中的默认视图,不改写 `skillfs-views.toml`。需要持久化分配时显式编辑配置,修改后重挂载生效。

需要长期保持挂载时改用 managed 模式:它启动一个 detached supervisor,worker 意外退出后 会自动重挂,不需要自己管理 pid file。

```bash skillfs mount ~/.openclaw/skills ~/.openclaw/skills --managed ```

---

### STEP 4:验证挂载状态

```bash # 检查进程是否存活 kill -0 $(cat /tmp/skillfs.pid) 2>/dev/null && echo "mounted" || echo "not mounted"

# 查看实时日志 tail -f /tmp/skillfs-$(cat /tmp/skillfs.pid).log ```

---

### STEP 5:卸载

```bash # managed 挂载:用 stop 停止 supervisor 并卸载 # 非 managed 或残留挂载:stop 也会直接卸载,可作为通用兜底 skillfs stop ~/.openclaw/skills

# 前台挂载也可以直接向进程发信号 kill -TERM $(cat /tmp/skillfs.pid) sleep 1

# 卸载后目录恢复完整内容 ls ~/.openclaw/skills/

# 若进程意外退出,手动清理残留挂载: fusermount3 -u ~/.openclaw/skills ```

PID 文件在卸载成功后自动删除。

---

## 关键说明

- SkillFS **不是只读**文件系统:挂载期间对 skill 的写入(含 install / update / remove) 会透传到底层 source 树,写 `SKILL.md` 还会触发重新解析 - 但 views.toml 是例外:挂载期间它对 FUSE 不可见,需修改时先卸载 - 不生成 views.toml 也可直接挂载,此时所有 skill 均在主视图中可见 - 该工作区不能同时作为 ws-ckpt 工作区根初始化 —— in-place 挂载会让 `rename(2)` 返回 `EBUSY`。先 `init` 再挂载,或先卸载再 `init`

---

## 故障排查

| 错误 | 解决方案 | |------|----------| | `Package fuse3 was not found` | `yum install -y fuse3 fuse3-devel` | | `Transport endpoint is not connected` | `fusermount3 -u <mountpoint>` 后重新挂载 | | skill 消失于列表 | 统一 skill 目录名与 views.toml 字符串(不是 SKILL.md 的 `name:`) | | skill-discover 不完整 | 重新运行 `skillfs classify` 或手动编辑 views.toml | | 原地挂载后无法写 views.toml | 先卸载,编辑,再重新挂载 |

Intended uses

  • Use Skillfs Mount when this documented workflow matches the task.

Implementation notes

  • Review the source instructions and adapt inputs before running the workflow.

Try it, then tell us how it went

A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.