// SCAN REPORT

Local Code

27 tools. 16 can modify or destroy data without limits.

16 write tools that can modify data. Rate limits recommended.

Last updated: 2 Jul 2026

16 can modify or destroy data

11 read-only

27 tools total

Community server · catalogue entry verified 02/07/2026

How to control Local Code ↓

TOOL MAP

What Local Code exposes to your agents

Read (11) Write / Execute (16) Destructive / Financial (0)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

High Risk

The most dangerous Local Code tools

Execute · High mcp_run_command_safely Command execution is inherently an Execute category risk. Execute · High mcp_run_compile_check This tool triggers execution of an external Python command (compileall module) whose effects depend on the project state and Python environment. Execute · High mcp_run_pytest The tool runs an external process (pytest) whose behavior depends entirely on the test suite's implementation. Write · Low mcp_generate_changelog_md This tool creates or modifies a changelog markdown file, which is a reversible write operation.

16 of Local Code's 27 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Local Code

PolicyLayer is an MCP gateway — it sits between your AI agents and Local Code, and nothing reaches the server without passing your rules. These are the rules we recommend:

Rate limit write operations

{
  "create_example_config": {
    "limits": [
      {
        "counter": "create_example_config_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "list_allowed_roots": {
    "limits": [
      {
        "counter": "list_allowed_roots_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

Create a free account and register Local Code — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON LOCAL CODE →

Instant setup, no code required.

FULL CATALOGUE

All 27 Local Code tools

EXECUTE 3 tools

Execute mcp_run_command_safely Run a command only if it matches the configured allowlist and project root policy. Execute mcp_run_compile_check Run python -m compileall . in an allowed project root. Execute mcp_run_pytest Run python -m pytest in an allowed project root.

WRITE 13 tools

Write create_example_config Create an example config.json file that allows C:/hwpmcp and C:/OfficeMCP. Write mcp_create_change_zip Create a ZIP containing selected changed files and generated markdown reports. Write mcp_create_full_snapshot_zip Create a full snapshot ZIP, excluding ignored directories. Write mcp_generate_changelog_md Generate a Korean markdown changelog/report inside an allowed project root. Write mcp_git_add Stage selected relative file paths with git add. Write mcp_git_commit Create a git commit from already staged changes. Write mcp_git_commit_files Stage selected relative file paths and create one git commit. Write mcp_git_create_branch Create a safety branch. Branch name must start with ai-change/. Write mcp_git_diagnose Diagnose Git execution from inside the MCP server process. Write mcp_git_restore_file Restore a changed file with git restore -- <file>. Write mcp_replace_in_file Replace exact text snippets in a file. replacements_json is a JSON list with old/new/replace_all. Write mcp_restore_from_backup Restore a file from a backup created by this program. Write mcp_write_file_safe Write a text file after creating a backup when the file already exists.

READ 11 tools

Read list_allowed_roots Return the configured project roots that this MCP server may access. Read mcp_detect_project_type Detect Python, Node, .NET, PowerShell, or MCP-like project markers. Read mcp_get_file_info Return size, type, and metadata for a path inside an allowed root. Read mcp_git_changed_files Return changed files from git status --porcelain. Read mcp_git_diff Run git diff or git diff --staged in an allowed project root. Read mcp_git_status Run git status --short in an allowed project root. Read mcp_list_project_files List files inside an allowed project root. Read mcp_read_file Read a text file inside an allowed project root. Read mcp_search_files Search file and directory paths by regex inside an allowed project root. Read mcp_search_text Search text in readable source files inside an allowed project root. Read mcp_summarize_project Summarize project type, file counts, extensions, and total scanned size.

EXPLORE

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

FAQ

Questions about Local Code

How do I prevent bulk modifications through Local Code? +

The Local Code server has 13 write tools including create_example_config, mcp_create_change_zip, mcp_create_full_snapshot_zip. Set a rate limit in your policy -- for example, 10 calls per hour prevents an agent from making more than 10 modifications per hour. PolicyLayer enforces this at the gateway, before calls reach Local Code.

How many tools does the Local Code MCP server expose? +

27 tools across 3 categories: Execute, Read, Write. 11 are read-only. 16 can modify, create, or delete data.

How do I enforce a policy on Local Code? +

Register the Local Code MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.

Enforce policy on every Local Code tool call.

Deterministic rules across all 27 Local Code tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON LOCAL CODE →

Instant setup, no code required.

27 Local Code tools catalogued and risk-classified — across an index of 43,000+ MCP servers.