regex-builder

name	regex-builder
description	Builds, explains, and tests regular expressions with step-by-step breakdowns
version	1.0.0
author	go-on-team
tags	["regex","regular-expression","text-processing","pattern-matching"]
min_go_on_version	1.0.0

Regex Builder Skill

Helps construct regular expressions from natural-language descriptions, explains what an existing regex does in plain English, and tests patterns against sample text. Supports PCRE, ECMAScript, Rust regex crate, Python re module, and Go regexp syntax variants.

How It Works

1. Describe to Regex — Converts natural language ("match an email address") into a working regex with explanation of each component
2. Regex to Explain — Parses an existing regex token by token and produces a plain-English breakdown of what it matches
3. Test — Runs the regex against provided sample strings and reports matches, captures, and positions
4. Optimize — Suggests simplifications, removes unnecessary escapes, converts capturing groups to non-capturing, and flags catastrophic backtracking patterns
5. Translate — Converts regex between different engine syntaxes (e.g., PCRE lookbehinds → JavaScript alternatives)

Input Schema

Parameter	Type	Description
description	string	Natural-language description (used when action=generate)
pattern	string	Existing regex pattern (used when action=explain or action=test)
test_strings	string[]	Optional: sample strings to test the pattern against
engine	string	Regex engine: rust, python, javascript, pcre, go (default: rust)
action	string	Action: generate, explain, test, optimize (default: generate)
flags	string	Optional: regex flags (i, m, s, x, combined)

Example

{
  "description": "Match a valid HTTP or HTTPS URL with optional path, query string, and fragment",
  "engine": "rust",
  "action": "generate",
  "test_strings": [
    "https://example.com/path?q=search#section",
    "http://localhost:8080",
    "ftp://invalid"
  ]
}

Example output (abbreviated):

// Generated pattern:
// ^https?://[\\w.-]+(?::\\d{1,5})?(?:/[\\w./-]*)*(?:\\?[\\w&=.-]*)?(?:#[\\w-]*)?$

// Breakdown:
// ^                     ─ Start of string
// https?                ─ Literal "http" + optional "s"
// ://                   ─ Literal "://"
// [\\w.-]+              ─ Domain name (word chars, dots, hyphens)
// (?::\\d{1,5})?        ─ Optional port number
// (?:/[\\w./-]*)*       ─ Optional path segments
// (?:\\?[\\w&=.-]*)?    ─ Optional query string
// (?:#[\\w-]*)?         ─ Optional fragment
// $                     ─ End of string

// Test results:
// ✅ https://example.com/path?q=search#section — MATCH
// ✅ http://localhost:8080 — MATCH
// ❌ ftp://invalid — NO MATCH (expected: wrong protocol)