mcpField
F
filesystem

filesystem

github.com/oimiragieoAgent workflows44729 installsUpdated 17w ago
Repository

About filesystem

filesystem is a skill that guides file operations with Claude Code's built-in tools. It covers finding and reading files, searching content, editing, verifying changes, and managing directories.

Key features

  1. 01Read one or multiple files
  2. 02Find files by pattern
  3. 03Search file contents
  4. 04Create and modify files
  5. 05Navigate and manage directories
  6. 06Read before editing and verify changes

Use cases

  1. 01Read multiple project files
  2. 02Find files by type or name
  3. 03Search code for a function or text
  4. 04Modify files and check the results

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/oimiragieo/agent-studio --skill filesystem

After installing, type/filesystemin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

commands/
filesystem.md
hooks/
post-execute.cjs
pre-execute.cjs
references/
research-requirements.md
rules/
filesystem.md
schemas/
input.schema.json
output.schema.json
scripts/
main.cjs
templates/
implementation-template.md

Details

From the project README

Filesystem Skill

Filesystem Skill - Guidance for file system operations using Claude Code's built-in tools. Read, write, search, and manage files efficiently. - Reading single or multiple files - Creating or modifying files - Searching for files by pattern - Searching file contents - Directory navigation

Claude Code Built-in Tools

Reading Files

Read Tool: Read file contents

Read file_path="/path/to/file.txt"

Options:

  • offset: Start line (for large files)
  • limit: Number of lines to read
Finding Files

Glob Tool: Find files by pattern

Glob pattern="**/*.ts"

Common patterns:

  • **/*.ts - All TypeScript files
  • src/**/*.tsx - React components in src
  • **/test*.js - Test files anywhere
Searching Content

Primary Search: Use pnpm search:code "query" or Skill({ skill: 'ripgrep' }) for content searching. The built-in Grep tool is fallback-only.

Grep Tool: Search file contents (fallback only — prefer pnpm search:code or ripgrep skill first)

Grep pattern="function myFunc" path="/src"

Options:

  • output_mode: "content", "files_with_matches", or "count"
  • -A, -B, -C: Context lines
Writing Files

Write Tool: Create or overwrite files

Write file_path="/path/to/file.txt" content="..."

Edit Tool: Modify existing files

Edit file_path="/path/to/file.txt" old_string="..." new_string="..."
Directory Operations

Bash Tool: For directory operations

# List directory
ls -la /path/to/dir

# Create directory
mkdir -p /path/to/new/dir

# Move/rename
mv /old/path /new/path

</execution_process>

<best_practices>

Common Workflows

Reading Multiple Files
# Read files in parallel (multiple Read calls in one message)
Read file_path="/src/app.ts"
Read file_path="/src/config.ts"
Read file_path="/src/utils.ts"
Search and Read
# 1. Find files
Glob pattern="**/*.config.ts"

# 2. Read matching files
Read file_path="/path/from/glob/result"
Find and Replace
# 1. Search for pattern
Grep pattern="oldFunction" path="/src"

# 2. Edit each file
Edit file_path="/src/file.ts" old_string="oldFunction" new_string="newFunction"

Best Practices

  1. Read Before Edit: Always read a file before editing it
  2. Use Glob Over Bash: Prefer Glob to find for file discovery
  3. Search Tool Hierarchy: Use pnpm search:code or Skill({ skill: 'ripgrep' }) first; use the Grep tool as fallback over raw bash grep only
  4. Parallel Reads: Read multiple files in one message for speed
  5. Verify Changes: Read file after editing to verify

</best_practices>

**Find and read all TypeScript files**:
Glob pattern="src/**/*.ts"
# Then read the results
Read file_path="/src/app.ts"

</usage_example>

<usage_example> Search for a function and edit it:

Grep pattern="export function oldName" path="/src"
# Found in /src/utils.ts:23

Edit file_path="/src/utils.ts" old_string="export function oldName" new_string="export function newName"

</usage_example>

Rules

  • Always read files before editing
  • Use built-in tools (Read, Glob, Grep) instead of bash equivalents
  • Verify changes after editing

Memory Protocol (MANDATORY)

Before starting:

cat .claude/context/memory/learnings.md

After completing:

  • New pattern -> .claude/context/memory/learnings.md
  • Issue found -> .claude/context/memory/issues.md
  • Decision made -> .claude/context/memory/decisions.md

ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.

Comments

Markdown supported