Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
MCP server that gives AI agents sandboxed access to local files and directories. Supports reading, writing, editing, moving, deleting, and searching files within explicitly allowed paths. Access controls block directory traversal and system file access at the server level.
Unlike direct filesystem access, it enforces directory sandboxing and path validation without requiring custom permission logic in each agent.
When to use it
- Reading project source files before generating code changes
- Scanning error logs to identify recurring failures
- Organizing downloaded files into categorized folders
- Writing generated documentation to a local docs directory
- Searching a codebase for all files matching a name pattern
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Analyze our app error logs and summarize the issues
- 1Lists files in the application log directory
- 2Reads the contents of the error log file
- 3Searches for files containing ERROR pattern across the log directory
- 4Generates a summary of identified issues and patterns
- 5Writes the summary report to a markdown file
Error summary written to /reports/error-summary.md with categorized issues and counts