Workspace Guardian

v1.0.9

Stop AI from scattering files everywhere. This skill enforces consistent naming and directory structure for all AI outputs — saves tokens, cuts management ti...

1· 82·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for lujun2508/file-rules.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Workspace Guardian" (lujun2508/file-rules) from ClawHub.
Skill page: https://clawhub.ai/lujun2508/file-rules
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install file-rules

ClawHub CLI

Package manager switcher

npx clawhub@latest install file-rules
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The name/description (enforce consistent naming and directory structure) matches the SKILL.md: rules for filenames, project directories, temp cleanup, and agent workspace management. There are no unrelated environment variables, binaries, or installs requested.
Instruction Scope
Instructions grant the agent authority to create, move, and delete files within agent-managed workspaces (defaulting to per-project directories and ~/.openclaw/workspace-*/). The skill instructs periodic checks and deletion of rejected/temporary files (with a prompt before bulk cleanup). It does not instruct reading or transmitting unrelated files or contacting external endpoints, but it does permit filesystem-modifying actions in agent workspaces, which is expected for this purpose and should be accepted consciously by the user.
Install Mechanism
No install spec and no code files — instruction-only skill. Nothing is written to disk by an installer.
Credentials
No environment variables, credentials, or external config paths are required. The lack of secrets or unrelated credentials is proportionate to the stated functionality.
Persistence & Privilege
always:false (default). The skill allows autonomous invocation (platform default). Because the instructions authorize deletion and cleanup within agent workspaces, granting autonomy means the agent could perform those filesystem operations during use — the user should be aware and require prompts/confirmations when desired.
Assessment
This skill is coherent with its stated goal and doesn't request secrets or installs, but it does instruct the agent to create, move, and delete files in the agent's workspaces (including ~/.openclaw/workspace-*/ and default drafts/output/temp locations). Before installing: 1) Decide whether you trust the agent to manage/delete files and ensure important data is backed up; 2) Prefer to explicitly specify project paths when you want files written somewhere specific; 3) Keep the agent configured to prompt before bulk cleanup or deletions if you want manual control; 4) Test the skill in a non-critical workspace first to confirm behavior. If you do not want any automatic cleanup, consider disabling autonomous invocation or requiring confirmations for maintenance actions.

Like a lobster shell, security has layers — review code before you run it.

Runtime requirements

OSWindows · Linux · macOS
latestvk97359tassj1ak58r805drbdsd84t52z
82downloads
1stars
9versions
Updated 1w ago
v1.0.9
MIT-0
Windows, Linux, macOS

Workspace Guardian

Unless the user explicitly specifies a path, this skill governs where and how files are created.


Core Rules

One Project = One Directory

All files belong to a project. Never scatter files in ~/, desktop, downloads, or timestamped folders.

Naming Standard

{date}_{description}.{ext}

Example: 2026-04-14_report.xlsx

Scripts: {verb}_{object}.py (e.g., generate_report.py)

Directory Structure

{project-root}/
├── output/     ← final deliverables
├── scripts/    ← scripts and tools
├── data/       ← data files
├── docs/       ← documentation
├── temp/       ← temporary files
└── README.md

Special Cases

User Specifies a Path

Follow the user's instruction.

No Clear Project

Output goes to the Agent's own workspace:

  • drafts/ — pending content
  • output/ — approved content
  • temp/ — temporary, clean after use

Agent Workspaces

Files in ~/.openclaw/workspace-*/ are managed by each Agent independently.


Periodic Maintenance

Scheduled Review

Periodically (every few days during normal usage), check for:

  • Old temp files that were never cleaned up
  • Multiple script versions that could be consolidated
  • Empty directories left behind

When found, ask the user: "发现 {issue},是否清理?" — wait for confirmation before taking action.

Do not主动清理大量文件 or run cleanup during busy times.

Maintenance

Rejected Outputs

When the user declines an output: remove the file promptly.

Temporary Files

Clean up temp/ after task completion. Old temp files should not accumulate.

Versioned Scripts

When updating a script, archive the old version with a date suffix (e.g., generate_report_v4_2026-04-09.py).

Empty Directories

Do not leave empty directories behind.


Self-Check

After creating a file, briefly confirm the location:

📁 Saved: {project-dir}/{filename}

Comments

Loading comments...