# castovdev/castov [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/castovdev/castov  
**GitHub Stars:** 1  
**npm Downloads (last month):** 1858  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/castovdev-castov

## Description
Deterministic local developer utilities for coding agents: inspect listening ports and processes, compute IPv4 CIDR subnet ranges, decode creation timestamps from UUIDv7, ULID, Snowflake and MongoDB ObjectId, and convert Unix epochs with zero telemetry.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "castov": {
    "command": "npx",
    "args": ["-y","castov"]
  }
}
```

## Documentation

## What castovdev/castov does

castovdev/castov provides local developer utilities through the Model Context Protocol and a command-line interface. It is intended for coding agents that need exact results for operations such as checking which process owns a listening port, calculating an IPv4 subnet, converting an epoch value, or identifying the creation time embedded in a supported ID.

The package also includes offline utilities for cron schedules, IANA time-zone conversion, cryptographic material generation, JWT inspection, digest calculation, and detecting leaked secrets in code or configuration text. Processing is described as local, with no external analytics, telemetry, or cloud round trips.

## How it works

The MCP entry point runs over JSON-RPC 2.0 using stdio. It supports both Content-Length framing and newline-delimited JSON stream messages. Agents invoke named tools, while the same underlying capabilities are available through the `castov` CLI.

The server exposes ten tools: `castov_port`, `castov_cidr`, `castov_detect_id`, `castov_timestamp`, `castov_cron`, `castov_timezone`, `castov_secret`, `castov_jwt_decode`, `castov_hash`, and `castov_scan_secrets`. Port inspection can identify listening services and process collisions; the identifier tool handles UUIDv7, ULID, Snowflake, and MongoDB ObjectId values.

## Setup and configuration

Run the MCP server without a global installation using:

```bash
npx -y castov mcp
```

For Claude Desktop, add a `castov` entry to the MCP configuration with `npx` as the command and `-y castov mcp` as its arguments. Cursor and Windsurf provide an MCP server settings interface where the command can be entered directly; Cursor also supports a project-level `.cursor/mcp.json` entry with the same command and arguments.

The CLI can be run with `npx castov --help`, installed globally with `npm install -g castov`, or installed through the documented Homebrew tap. A Docker image is also available and runs the command-line tool in an Alpine-based Node.js container under the non-root `node` user.

## Tools and capabilities

- `castov_port` inspects listening ports, maps protocols, and examines process collisions.
- `castov_cidr` calculates IPv4 network and wildcard masks, broadcast addresses, and usable host counts.
- `castov_detect_id` extracts creation timestamps from UUIDv7, ULID, Snowflake, and MongoDB ObjectId values.
- `castov_timestamp` converts Unix epochs to ISO 8601, UTC, and localized representations in both directions.
- `castov_cron` explains five-field cron schedules and calculates the next five runs.
- `castov_timezone` converts times between IANA zones and reports DST transition warnings.
- `castov_secret` generates 256-bit hexadecimal or Base64URL values, API keys, and sortable UUIDv7 secrets.
- `castov_jwt_decode` decodes JWT headers and payloads and checks expiration status without network calls.
- `castov_hash` computes SHA-1, SHA-256, SHA-384, SHA-512, MD5, Base64, and Base64URL encodings.
- `castov_scan_secrets` checks snippets, git diffs, and `.env` blocks for exposed API keys and private keys.

## Limitations and notes

The MCP interface is designed for local execution and exact utility operations rather than remote service access. The documented process and port command can terminate a process when invoked with its CLI `--kill` option, while the MCP tool list describes the agent-facing tools as read-only. Secret scanning recognizes API-key and private-key leakage patterns, so its results should be treated as an audit aid rather than a replacement for repository security controls.

_Full upstream README: https://allmcps.com/mcp/castovdev-castov/readme_