Add comprehensive development roadmap via GitHub Issues

Created 10 detailed GitHub issues covering:
- Project activation and management UI (#1-2)
- Worker node coordination and visualization (#3-4)
- Automated GitHub repository scanning (#5)
- Intelligent model-to-issue matching (#6)
- Multi-model task execution system (#7)
- N8N workflow integration (#8)
- Hive-Bzzz P2P bridge (#9)
- Peer assistance protocol (#10)

Each issue includes detailed specifications, acceptance criteria,
technical implementation notes, and dependency mapping.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>

mcp-server/node_modules/markdown-it/LICENSE

@@ -0,0 +1,22 @@
+Copyright (c) 2014 Vitaly Puzrin, Alex Kocharin.
+
+Permission is hereby granted, free of charge, to any person
+obtaining a copy of this software and associated documentation
+files (the "Software"), to deal in the Software without
+restriction, including without limitation the rights to use,
+copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the
+Software is furnished to do so, subject to the following
+conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
+OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+OTHER DEALINGS IN THE SOFTWARE.

mcp-server/node_modules/markdown-it/README.md

@@ -0,0 +1,324 @@
+# markdown-it <!-- omit in toc -->
+
+[![CI](https://github.com/markdown-it/markdown-it/actions/workflows/ci.yml/badge.svg)](https://github.com/markdown-it/markdown-it/actions/workflows/ci.yml)
+[![NPM version](https://img.shields.io/npm/v/markdown-it.svg?style=flat)](https://www.npmjs.org/package/markdown-it)
+[![Coverage Status](https://coveralls.io/repos/markdown-it/markdown-it/badge.svg?branch=master&service=github)](https://coveralls.io/github/markdown-it/markdown-it?branch=master)
+[![Gitter](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/markdown-it/markdown-it)
+
+> Markdown parser done right. Fast and easy to extend.
+
+__[Live demo](https://markdown-it.github.io)__
+
+- Follows the __[CommonMark spec](http://spec.commonmark.org/)__ + adds syntax extensions & sugar (URL autolinking, typographer).
+- Configurable syntax! You can add new rules and even replace existing ones.
+- High speed.
+- [Safe](https://github.com/markdown-it/markdown-it/tree/master/docs/security.md) by default.
+- Community-written __[plugins](https://www.npmjs.org/browse/keyword/markdown-it-plugin)__ and [other packages](https://www.npmjs.org/browse/keyword/markdown-it) on npm.
+
+__Table of content__
+
+- [Install](#install)
+- [Usage examples](#usage-examples)
+  - [Simple](#simple)
+  - [Init with presets and options](#init-with-presets-and-options)
+  - [Plugins load](#plugins-load)
+  - [Syntax highlighting](#syntax-highlighting)
+  - [Linkify](#linkify)
+- [API](#api)
+- [Syntax extensions](#syntax-extensions)
+  - [Manage rules](#manage-rules)
+- [Benchmark](#benchmark)
+- [markdown-it for enterprise](#markdown-it-for-enterprise)
+- [Authors](#authors)
+- [References / Thanks](#references--thanks)
+
+## Install
+
+**node.js**:
+
+```bash
+npm install markdown-it
+```
+
+**browser (CDN):**
+
+- [jsDeliver CDN](http://www.jsdelivr.com/#!markdown-it "jsDelivr CDN")
+- [cdnjs.com CDN](https://cdnjs.com/libraries/markdown-it "cdnjs.com")
+
+
+## Usage examples
+
+See also:
+
+- __[API documentation](https://markdown-it.github.io/markdown-it/)__ - for more
+  info and examples.
+- [Development info](https://github.com/markdown-it/markdown-it/tree/master/docs) -
+  for plugins writers.
+
+
+### Simple
+
+```js
+// node.js
+// can use `require('markdown-it')` for CJS
+import markdownit from 'markdown-it'
+const md = markdownit()
+const result = md.render('# markdown-it rulezz!');
+
+// browser with UMD build, added to "window" on script load
+// Note, there is no dash in "markdownit".
+const md = window.markdownit();
+const result = md.render('# markdown-it rulezz!');
+```
+
+Single line rendering, without paragraph wrap:
+
+```js
+import markdownit from 'markdown-it'
+const md = markdownit()
+const result = md.renderInline('__markdown-it__ rulezz!');
+```
+
+
+### Init with presets and options
+
+(*) presets define combinations of active rules and options. Can be
+`"commonmark"`, `"zero"` or `"default"` (if skipped). See
+[API docs](https://markdown-it.github.io/markdown-it/#MarkdownIt.new) for more details.
+
+```js
+import markdownit from 'markdown-it'
+
+// commonmark mode
+const md = markdownit('commonmark')
+
+// default mode
+const md = markdownit()
+
+// enable everything
+const md = markdownit({
+  html: true,
+  linkify: true,
+  typographer: true
+})
+
+// full options list (defaults)
+const md = markdownit({
+  // Enable HTML tags in source
+  html:         false,
+
+  // Use '/' to close single tags (<br />).
+  // This is only for full CommonMark compatibility.
+  xhtmlOut:     false,
+
+  // Convert '\n' in paragraphs into <br>
+  breaks:       false,
+
+  // CSS language prefix for fenced blocks. Can be
+  // useful for external highlighters.
+  langPrefix:   'language-',
+
+  // Autoconvert URL-like text to links
+  linkify:      false,
+
+  // Enable some language-neutral replacement + quotes beautification
+  // For the full list of replacements, see https://github.com/markdown-it/markdown-it/blob/master/lib/rules_core/replacements.mjs
+  typographer:  false,
+
+  // Double + single quotes replacement pairs, when typographer enabled,
+  // and smartquotes on. Could be either a String or an Array.
+  //
+  // For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+  // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+  quotes: '“”‘’',
+
+  // Highlighter function. Should return escaped HTML,
+  // or '' if the source string is not changed and should be escaped externally.
+  // If result starts with <pre... internal wrapper is skipped.
+  highlight: function (/*str, lang*/) { return ''; }
+});
+```
+
+### Plugins load
+
+```js
+import markdownit from 'markdown-it'
+
+const md = markdownit
+  .use(plugin1)
+  .use(plugin2, opts, ...)
+  .use(plugin3);
+```
+
+
+### Syntax highlighting
+
+Apply syntax highlighting to fenced code blocks with the `highlight` option:
+
+```js
+import markdownit from 'markdown-it'
+import hljs from 'highlight.js' // https://highlightjs.org
+
+// Actual default values
+const md = markdownit({
+  highlight: function (str, lang) {
+    if (lang && hljs.getLanguage(lang)) {
+      try {
+        return hljs.highlight(str, { language: lang }).value;
+      } catch (__) {}
+    }
+
+    return ''; // use external default escaping
+  }
+});
+```
+
+Or with full wrapper override (if you need assign class to `<pre>` or `<code>`):
+
+```js
+import markdownit from 'markdown-it'
+import hljs from 'highlight.js' // https://highlightjs.org
+
+// Actual default values
+const md = markdownit({
+  highlight: function (str, lang) {
+    if (lang && hljs.getLanguage(lang)) {
+      try {
+        return '<pre><code class="hljs">' +
+               hljs.highlight(str, { language: lang, ignoreIllegals: true }).value +
+               '</code></pre>';
+      } catch (__) {}
+    }
+
+    return '<pre><code class="hljs">' + md.utils.escapeHtml(str) + '</code></pre>';
+  }
+});
+```
+
+### Linkify
+
+`linkify: true` uses [linkify-it](https://github.com/markdown-it/linkify-it). To
+configure linkify-it, access the linkify instance through `md.linkify`:
+
+```js
+md.linkify.set({ fuzzyEmail: false });  // disables converting email to link
+```
+
+
+## API
+
+__[API documentation](https://markdown-it.github.io/markdown-it/)__
+
+If you are going to write plugins, please take a look at
+[Development info](https://github.com/markdown-it/markdown-it/tree/master/docs).
+
+
+## Syntax extensions
+
+Embedded (enabled by default):
+
+- [Tables](https://help.github.com/articles/organizing-information-with-tables/) (GFM)
+- [Strikethrough](https://help.github.com/articles/basic-writing-and-formatting-syntax/#styling-text) (GFM)
+
+Via plugins:
+
+- [subscript](https://github.com/markdown-it/markdown-it-sub)
+- [superscript](https://github.com/markdown-it/markdown-it-sup)
+- [footnote](https://github.com/markdown-it/markdown-it-footnote)
+- [definition list](https://github.com/markdown-it/markdown-it-deflist)
+- [abbreviation](https://github.com/markdown-it/markdown-it-abbr)
+- [emoji](https://github.com/markdown-it/markdown-it-emoji)
+- [custom container](https://github.com/markdown-it/markdown-it-container)
+- [insert](https://github.com/markdown-it/markdown-it-ins)
+- [mark](https://github.com/markdown-it/markdown-it-mark)
+- ... and [others](https://www.npmjs.org/browse/keyword/markdown-it-plugin)
+
+
+### Manage rules
+
+By default all rules are enabled, but can be restricted by options. On plugin
+load all its rules are enabled automatically.
+
+```js
+import markdownit from 'markdown-it'
+
+// Activate/deactivate rules, with currying
+const md = markdownit()
+  .disable(['link', 'image'])
+  .enable(['link'])
+  .enable('image');
+
+// Enable everything
+const md = markdownit({
+  html: true,
+  linkify: true,
+  typographer: true,
+});
+```
+
+You can find all rules in sources:
+
+- [`parser_core.mjs`](lib/parser_core.mjs)
+- [`parser_block.mjs`](lib/parser_block.mjs)
+- [`parser_inline.mjs`](lib/parser_inline.mjs)
+
+
+## Benchmark
+
+Here is the result of readme parse at MB Pro Retina 2013 (2.4 GHz):
+
+```bash
+npm run benchmark-deps
+benchmark/benchmark.mjs readme
+
+Selected samples: (1 of 28)
+ > README
+
+Sample: README.md (7774 bytes)
+ > commonmark-reference x 1,222 ops/sec ±0.96% (97 runs sampled)
+ > current x 743 ops/sec ±0.84% (97 runs sampled)
+ > current-commonmark x 1,568 ops/sec ±0.84% (98 runs sampled)
+ > marked x 1,587 ops/sec ±4.31% (93 runs sampled)
+```
+
+__Note.__ CommonMark version runs with [simplified link normalizers](https://github.com/markdown-it/markdown-it/blob/master/benchmark/implementations/current-commonmark/index.mjs)
+for more "honest" compare. Difference is ≈1.5×.
+
+As you can see, `markdown-it` doesn't pay with speed for its flexibility.
+Slowdown of "full" version caused by additional features not available in
+other implementations.
+
+
+## markdown-it for enterprise
+
+Available as part of the Tidelift Subscription.
+
+The maintainers of `markdown-it` and thousands of other packages are working with Tidelift to deliver commercial support and maintenance for the open source dependencies you use to build your applications. Save time, reduce risk, and improve code health, while paying the maintainers of the exact dependencies you use. [Learn more.](https://tidelift.com/subscription/pkg/npm-markdown-it?utm_source=npm-markdown-it&utm_medium=referral&utm_campaign=enterprise&utm_term=repo)
+
+
+## Authors
+
+- Alex Kocharin [github/rlidwka](https://github.com/rlidwka)
+- Vitaly Puzrin [github/puzrin](https://github.com/puzrin)
+
+_markdown-it_ is the result of the decision of the authors who contributed to
+99% of the _Remarkable_ code to move to a project with the same authorship but
+new leadership (Vitaly and Alex). It's not a fork.
+
+## References / Thanks
+
+Big thanks to [John MacFarlane](https://github.com/jgm) for his work on the
+CommonMark spec and reference implementations. His work saved us a lot of time
+during this project's development.
+
+**Related Links:**
+
+- https://github.com/jgm/CommonMark - reference CommonMark implementations in C & JS,
+  also contains latest spec & online demo.
+- http://talk.commonmark.org - CommonMark forum, good place to collaborate
+  developers' efforts.
+
+**Ports**
+
+- [motion-markdown-it](https://github.com/digitalmoksha/motion-markdown-it) - Ruby/RubyMotion
+- [markdown-it-py](https://github.com/ExecutableBookProject/markdown-it-py)- Python

mcp-server/node_modules/markdown-it/bin/markdown-it.mjs

@@ -0,0 +1,107 @@
+#!/usr/bin/env node
+/* eslint no-console:0 */
+
+import fs from 'node:fs'
+import argparse from 'argparse'
+import markdownit from '../index.mjs'
+
+const cli = new argparse.ArgumentParser({
+  prog: 'markdown-it',
+  add_help: true
+})
+
+cli.add_argument('-v', '--version', {
+  action: 'version',
+  version: JSON.parse(fs.readFileSync(new URL('../package.json', import.meta.url))).version
+})
+
+cli.add_argument('--no-html', {
+  help: 'Disable embedded HTML',
+  action: 'store_true'
+})
+
+cli.add_argument('-l', '--linkify', {
+  help: 'Autolink text',
+  action: 'store_true'
+})
+
+cli.add_argument('-t', '--typographer', {
+  help: 'Enable smartquotes and other typographic replacements',
+  action: 'store_true'
+})
+
+cli.add_argument('--trace', {
+  help: 'Show stack trace on error',
+  action: 'store_true'
+})
+
+cli.add_argument('file', {
+  help: 'File to read',
+  nargs: '?',
+  default: '-'
+})
+
+cli.add_argument('-o', '--output', {
+  help: 'File to write',
+  default: '-'
+})
+
+const options = cli.parse_args()
+
+function readFile (filename, encoding, callback) {
+  if (options.file === '-') {
+    // read from stdin
+    const chunks = []
+
+    process.stdin.on('data', function (chunk) { chunks.push(chunk) })
+
+    process.stdin.on('end', function () {
+      return callback(null, Buffer.concat(chunks).toString(encoding))
+    })
+  } else {
+    fs.readFile(filename, encoding, callback)
+  }
+}
+
+readFile(options.file, 'utf8', function (err, input) {
+  let output
+
+  if (err) {
+    if (err.code === 'ENOENT') {
+      console.error('File not found: ' + options.file)
+      process.exit(2)
+    }
+
+    console.error(
+      (options.trace && err.stack) ||
+      err.message ||
+      String(err))
+
+    process.exit(1)
+  }
+
+  const md = markdownit({
+    html: !options.no_html,
+    xhtmlOut: false,
+    typographer: options.typographer,
+    linkify: options.linkify
+  })
+
+  try {
+    output = md.render(input)
+  } catch (e) {
+    console.error(
+      (options.trace && e.stack) ||
+      e.message ||
+      String(e))
+
+    process.exit(1)
+  }
+
+  if (options.output === '-') {
+    // write to stdout
+    process.stdout.write(output)
+  } else {
+    fs.writeFileSync(options.output, output)
+  }
+})

mcp-server/node_modules/markdown-it/index.mjs

@@ -0,0 +1 @@
+export { default } from './lib/index.mjs'

mcp-server/node_modules/markdown-it/lib/common/html_blocks.mjs

@@ -0,0 +1,67 @@
+// List of valid html blocks names, according to commonmark spec
+// https://spec.commonmark.org/0.30/#html-blocks
+
+export default [
+  'address',
+  'article',
+  'aside',
+  'base',
+  'basefont',
+  'blockquote',
+  'body',
+  'caption',
+  'center',
+  'col',
+  'colgroup',
+  'dd',
+  'details',
+  'dialog',
+  'dir',
+  'div',
+  'dl',
+  'dt',
+  'fieldset',
+  'figcaption',
+  'figure',
+  'footer',
+  'form',
+  'frame',
+  'frameset',
+  'h1',
+  'h2',
+  'h3',
+  'h4',
+  'h5',
+  'h6',
+  'head',
+  'header',
+  'hr',
+  'html',
+  'iframe',
+  'legend',
+  'li',
+  'link',
+  'main',
+  'menu',
+  'menuitem',
+  'nav',
+  'noframes',
+  'ol',
+  'optgroup',
+  'option',
+  'p',
+  'param',
+  'search',
+  'section',
+  'summary',
+  'table',
+  'tbody',
+  'td',
+  'tfoot',
+  'th',
+  'thead',
+  'title',
+  'tr',
+  'track',
+  'ul'
+]

mcp-server/node_modules/markdown-it/lib/common/html_re.mjs

@@ -0,0 +1,25 @@
+// Regexps to match html elements
+
+const attr_name     = '[a-zA-Z_:][a-zA-Z0-9:._-]*'
+
+const unquoted      = '[^"\'=<>`\\x00-\\x20]+'
+const single_quoted = "'[^']*'"
+const double_quoted = '"[^"]*"'
+
+const attr_value  = '(?:' + unquoted + '|' + single_quoted + '|' + double_quoted + ')'
+
+const attribute   = '(?:\\s+' + attr_name + '(?:\\s*=\\s*' + attr_value + ')?)'
+
+const open_tag    = '<[A-Za-z][A-Za-z0-9\\-]*' + attribute + '*\\s*\\/?>'
+
+const close_tag   = '<\\/[A-Za-z][A-Za-z0-9\\-]*\\s*>'
+const comment     = '<!---?>|<!--(?:[^-]|-[^-]|--[^>])*-->'
+const processing  = '<[?][\\s\\S]*?[?]>'
+const declaration = '<![A-Za-z][^>]*>'
+const cdata       = '<!\\[CDATA\\[[\\s\\S]*?\\]\\]>'
+
+const HTML_TAG_RE = new RegExp('^(?:' + open_tag + '|' + close_tag + '|' + comment +
+                        '|' + processing + '|' + declaration + '|' + cdata + ')')
+const HTML_OPEN_CLOSE_TAG_RE = new RegExp('^(?:' + open_tag + '|' + close_tag + ')')
+
+export { HTML_TAG_RE, HTML_OPEN_CLOSE_TAG_RE }

mcp-server/node_modules/markdown-it/lib/common/utils.mjs

@@ -0,0 +1,304 @@
+// Utilities
+//
+
+import * as mdurl from 'mdurl'
+import * as ucmicro from 'uc.micro'
+import { decodeHTML } from 'entities'
+
+function _class (obj) { return Object.prototype.toString.call(obj) }
+
+function isString (obj) { return _class(obj) === '[object String]' }
+
+const _hasOwnProperty = Object.prototype.hasOwnProperty
+
+function has (object, key) {
+  return _hasOwnProperty.call(object, key)
+}
+
+// Merge objects
+//
+function assign (obj /* from1, from2, from3, ... */) {
+  const sources = Array.prototype.slice.call(arguments, 1)
+
+  sources.forEach(function (source) {
+    if (!source) { return }
+
+    if (typeof source !== 'object') {
+      throw new TypeError(source + 'must be object')
+    }
+
+    Object.keys(source).forEach(function (key) {
+      obj[key] = source[key]
+    })
+  })
+
+  return obj
+}
+
+// Remove element from array and put another array at those position.
+// Useful for some operations with tokens
+function arrayReplaceAt (src, pos, newElements) {
+  return [].concat(src.slice(0, pos), newElements, src.slice(pos + 1))
+}
+
+function isValidEntityCode (c) {
+  /* eslint no-bitwise:0 */
+  // broken sequence
+  if (c >= 0xD800 && c <= 0xDFFF) { return false }
+  // never used
+  if (c >= 0xFDD0 && c <= 0xFDEF) { return false }
+  if ((c & 0xFFFF) === 0xFFFF || (c & 0xFFFF) === 0xFFFE) { return false }
+  // control codes
+  if (c >= 0x00 && c <= 0x08) { return false }
+  if (c === 0x0B) { return false }
+  if (c >= 0x0E && c <= 0x1F) { return false }
+  if (c >= 0x7F && c <= 0x9F) { return false }
+  // out of range
+  if (c > 0x10FFFF) { return false }
+  return true
+}
+
+function fromCodePoint (c) {
+  /* eslint no-bitwise:0 */
+  if (c > 0xffff) {
+    c -= 0x10000
+    const surrogate1 = 0xd800 + (c >> 10)
+    const surrogate2 = 0xdc00 + (c & 0x3ff)
+
+    return String.fromCharCode(surrogate1, surrogate2)
+  }
+  return String.fromCharCode(c)
+}
+
+const UNESCAPE_MD_RE  = /\\([!"#$%&'()*+,\-./:;<=>?@[\\\]^_`{|}~])/g
+const ENTITY_RE       = /&([a-z#][a-z0-9]{1,31});/gi
+const UNESCAPE_ALL_RE = new RegExp(UNESCAPE_MD_RE.source + '|' + ENTITY_RE.source, 'gi')
+
+const DIGITAL_ENTITY_TEST_RE = /^#((?:x[a-f0-9]{1,8}|[0-9]{1,8}))$/i
+
+function replaceEntityPattern (match, name) {
+  if (name.charCodeAt(0) === 0x23/* # */ && DIGITAL_ENTITY_TEST_RE.test(name)) {
+    const code = name[1].toLowerCase() === 'x'
+      ? parseInt(name.slice(2), 16)
+      : parseInt(name.slice(1), 10)
+
+    if (isValidEntityCode(code)) {
+      return fromCodePoint(code)
+    }
+
+    return match
+  }
+
+  const decoded = decodeHTML(match)
+  if (decoded !== match) {
+    return decoded
+  }
+
+  return match
+}
+
+/* function replaceEntities(str) {
+  if (str.indexOf('&') < 0) { return str; }
+
+  return str.replace(ENTITY_RE, replaceEntityPattern);
+} */
+
+function unescapeMd (str) {
+  if (str.indexOf('\\') < 0) { return str }
+  return str.replace(UNESCAPE_MD_RE, '$1')
+}
+
+function unescapeAll (str) {
+  if (str.indexOf('\\') < 0 && str.indexOf('&') < 0) { return str }
+
+  return str.replace(UNESCAPE_ALL_RE, function (match, escaped, entity) {
+    if (escaped) { return escaped }
+    return replaceEntityPattern(match, entity)
+  })
+}
+
+const HTML_ESCAPE_TEST_RE = /[&<>"]/
+const HTML_ESCAPE_REPLACE_RE = /[&<>"]/g
+const HTML_REPLACEMENTS = {
+  '&': '&amp;',
+  '<': '&lt;',
+  '>': '&gt;',
+  '"': '&quot;'
+}
+
+function replaceUnsafeChar (ch) {
+  return HTML_REPLACEMENTS[ch]
+}
+
+function escapeHtml (str) {
+  if (HTML_ESCAPE_TEST_RE.test(str)) {
+    return str.replace(HTML_ESCAPE_REPLACE_RE, replaceUnsafeChar)
+  }
+  return str
+}
+
+const REGEXP_ESCAPE_RE = /[.?*+^$[\]\\(){}|-]/g
+
+function escapeRE (str) {
+  return str.replace(REGEXP_ESCAPE_RE, '\\$&')
+}
+
+function isSpace (code) {
+  switch (code) {
+    case 0x09:
+    case 0x20:
+      return true
+  }
+  return false
+}
+
+// Zs (unicode class) || [\t\f\v\r\n]
+function isWhiteSpace (code) {
+  if (code >= 0x2000 && code <= 0x200A) { return true }
+  switch (code) {
+    case 0x09: // \t
+    case 0x0A: // \n
+    case 0x0B: // \v
+    case 0x0C: // \f
+    case 0x0D: // \r
+    case 0x20:
+    case 0xA0:
+    case 0x1680:
+    case 0x202F:
+    case 0x205F:
+    case 0x3000:
+      return true
+  }
+  return false
+}
+
+/* eslint-disable max-len */
+
+// Currently without astral characters support.
+function isPunctChar (ch) {
+  return ucmicro.P.test(ch) || ucmicro.S.test(ch)
+}
+
+// Markdown ASCII punctuation characters.
+//
+// !, ", #, $, %, &, ', (, ), *, +, ,, -, ., /, :, ;, <, =, >, ?, @, [, \, ], ^, _, `, {, |, }, or ~
+// http://spec.commonmark.org/0.15/#ascii-punctuation-character
+//
+// Don't confuse with unicode punctuation !!! It lacks some chars in ascii range.
+//
+function isMdAsciiPunct (ch) {
+  switch (ch) {
+    case 0x21/* ! */:
+    case 0x22/* " */:
+    case 0x23/* # */:
+    case 0x24/* $ */:
+    case 0x25/* % */:
+    case 0x26/* & */:
+    case 0x27/* ' */:
+    case 0x28/* ( */:
+    case 0x29/* ) */:
+    case 0x2A/* * */:
+    case 0x2B/* + */:
+    case 0x2C/* , */:
+    case 0x2D/* - */:
+    case 0x2E/* . */:
+    case 0x2F/* / */:
+    case 0x3A/* : */:
+    case 0x3B/* ; */:
+    case 0x3C/* < */:
+    case 0x3D/* = */:
+    case 0x3E/* > */:
+    case 0x3F/* ? */:
+    case 0x40/* @ */:
+    case 0x5B/* [ */:
+    case 0x5C/* \ */:
+    case 0x5D/* ] */:
+    case 0x5E/* ^ */:
+    case 0x5F/* _ */:
+    case 0x60/* ` */:
+    case 0x7B/* { */:
+    case 0x7C/* | */:
+    case 0x7D/* } */:
+    case 0x7E/* ~ */:
+      return true
+    default:
+      return false
+  }
+}
+
+// Hepler to unify [reference labels].
+//
+function normalizeReference (str) {
+  // Trim and collapse whitespace
+  //
+  str = str.trim().replace(/\s+/g, ' ')
+
+  // In node v10 'ẞ'.toLowerCase() === 'Ṿ', which is presumed to be a bug
+  // fixed in v12 (couldn't find any details).
+  //
+  // So treat this one as a special case
+  // (remove this when node v10 is no longer supported).
+  //
+  if ('ẞ'.toLowerCase() === 'Ṿ') {
+    str = str.replace(/ẞ/g, 'ß')
+  }
+
+  // .toLowerCase().toUpperCase() should get rid of all differences
+  // between letter variants.
+  //
+  // Simple .toLowerCase() doesn't normalize 125 code points correctly,
+  // and .toUpperCase doesn't normalize 6 of them (list of exceptions:
+  // İ, ϴ, ẞ, Ω, K, Å - those are already uppercased, but have differently
+  // uppercased versions).
+  //
+  // Here's an example showing how it happens. Lets take greek letter omega:
+  // uppercase U+0398 (Θ), U+03f4 (ϴ) and lowercase U+03b8 (θ), U+03d1 (ϑ)
+  //
+  // Unicode entries:
+  // 0398;GREEK CAPITAL LETTER THETA;Lu;0;L;;;;;N;;;;03B8;
+  // 03B8;GREEK SMALL LETTER THETA;Ll;0;L;;;;;N;;;0398;;0398
+  // 03D1;GREEK THETA SYMBOL;Ll;0;L;<compat> 03B8;;;;N;GREEK SMALL LETTER SCRIPT THETA;;0398;;0398
+  // 03F4;GREEK CAPITAL THETA SYMBOL;Lu;0;L;<compat> 0398;;;;N;;;;03B8;
+  //
+  // Case-insensitive comparison should treat all of them as equivalent.
+  //
+  // But .toLowerCase() doesn't change ϑ (it's already lowercase),
+  // and .toUpperCase() doesn't change ϴ (already uppercase).
+  //
+  // Applying first lower then upper case normalizes any character:
+  // '\u0398\u03f4\u03b8\u03d1'.toLowerCase().toUpperCase() === '\u0398\u0398\u0398\u0398'
+  //
+  // Note: this is equivalent to unicode case folding; unicode normalization
+  // is a different step that is not required here.
+  //
+  // Final result should be uppercased, because it's later stored in an object
+  // (this avoid a conflict with Object.prototype members,
+  // most notably, `__proto__`)
+  //
+  return str.toLowerCase().toUpperCase()
+}
+
+// Re-export libraries commonly used in both markdown-it and its plugins,
+// so plugins won't have to depend on them explicitly, which reduces their
+// bundled size (e.g. a browser build).
+//
+const lib = { mdurl, ucmicro }
+
+export {
+  lib,
+  assign,
+  isString,
+  has,
+  unescapeMd,
+  unescapeAll,
+  isValidEntityCode,
+  fromCodePoint,
+  escapeHtml,
+  arrayReplaceAt,
+  isSpace,
+  isWhiteSpace,
+  isMdAsciiPunct,
+  isPunctChar,
+  escapeRE,
+  normalizeReference
+}

mcp-server/node_modules/markdown-it/lib/helpers/index.mjs

@@ -0,0 +1,11 @@
+// Just a shortcut for bulk export
+
+import parseLinkLabel from './parse_link_label.mjs'
+import parseLinkDestination from './parse_link_destination.mjs'
+import parseLinkTitle from './parse_link_title.mjs'
+
+export {
+  parseLinkLabel,
+  parseLinkDestination,
+  parseLinkTitle
+}

mcp-server/node_modules/markdown-it/lib/helpers/parse_link_destination.mjs

@@ -0,0 +1,77 @@
+// Parse link destination
+//
+
+import { unescapeAll } from '../common/utils.mjs'
+
+export default function parseLinkDestination (str, start, max) {
+  let code
+  let pos = start
+
+  const result = {
+    ok: false,
+    pos: 0,
+    str: ''
+  }
+
+  if (str.charCodeAt(pos) === 0x3C /* < */) {
+    pos++
+    while (pos < max) {
+      code = str.charCodeAt(pos)
+      if (code === 0x0A /* \n */) { return result }
+      if (code === 0x3C /* < */) { return result }
+      if (code === 0x3E /* > */) {
+        result.pos = pos + 1
+        result.str = unescapeAll(str.slice(start + 1, pos))
+        result.ok = true
+        return result
+      }
+      if (code === 0x5C /* \ */ && pos + 1 < max) {
+        pos += 2
+        continue
+      }
+
+      pos++
+    }
+
+    // no closing '>'
+    return result
+  }
+
+  // this should be ... } else { ... branch
+
+  let level = 0
+  while (pos < max) {
+    code = str.charCodeAt(pos)
+
+    if (code === 0x20) { break }
+
+    // ascii control characters
+    if (code < 0x20 || code === 0x7F) { break }
+
+    if (code === 0x5C /* \ */ && pos + 1 < max) {
+      if (str.charCodeAt(pos + 1) === 0x20) { break }
+      pos += 2
+      continue
+    }
+
+    if (code === 0x28 /* ( */) {
+      level++
+      if (level > 32) { return result }
+    }
+
+    if (code === 0x29 /* ) */) {
+      if (level === 0) { break }
+      level--
+    }
+
+    pos++
+  }
+
+  if (start === pos) { return result }
+  if (level !== 0) { return result }
+
+  result.str = unescapeAll(str.slice(start, pos))
+  result.pos = pos
+  result.ok = true
+  return result
+}

mcp-server/node_modules/markdown-it/lib/helpers/parse_link_label.mjs

@@ -0,0 +1,49 @@
+// Parse link label
+//
+// this function assumes that first character ("[") already matches;
+// returns the end of the label
+//
+
+export default function parseLinkLabel (state, start, disableNested) {
+  let level, found, marker, prevPos
+
+  const max = state.posMax
+  const oldPos = state.pos
+
+  state.pos = start + 1
+  level = 1
+
+  while (state.pos < max) {
+    marker = state.src.charCodeAt(state.pos)
+    if (marker === 0x5D /* ] */) {
+      level--
+      if (level === 0) {
+        found = true
+        break
+      }
+    }
+
+    prevPos = state.pos
+    state.md.inline.skipToken(state)
+    if (marker === 0x5B /* [ */) {
+      if (prevPos === state.pos - 1) {
+        // increase level if we find text `[`, which is not a part of any token
+        level++
+      } else if (disableNested) {
+        state.pos = oldPos
+        return -1
+      }
+    }
+  }
+
+  let labelEnd = -1
+
+  if (found) {
+    labelEnd = state.pos
+  }
+
+  // restore old state
+  state.pos = oldPos
+
+  return labelEnd
+}

mcp-server/node_modules/markdown-it/lib/helpers/parse_link_title.mjs

@@ -0,0 +1,66 @@
+// Parse link title
+//
+
+import { unescapeAll } from '../common/utils.mjs'
+
+// Parse link title within `str` in [start, max] range,
+// or continue previous parsing if `prev_state` is defined (equal to result of last execution).
+//
+export default function parseLinkTitle (str, start, max, prev_state) {
+  let code
+  let pos = start
+
+  const state = {
+    // if `true`, this is a valid link title
+    ok: false,
+    // if `true`, this link can be continued on the next line
+    can_continue: false,
+    // if `ok`, it's the position of the first character after the closing marker
+    pos: 0,
+    // if `ok`, it's the unescaped title
+    str: '',
+    // expected closing marker character code
+    marker: 0
+  }
+
+  if (prev_state) {
+    // this is a continuation of a previous parseLinkTitle call on the next line,
+    // used in reference links only
+    state.str = prev_state.str
+    state.marker = prev_state.marker
+  } else {
+    if (pos >= max) { return state }
+
+    let marker = str.charCodeAt(pos)
+    if (marker !== 0x22 /* " */ && marker !== 0x27 /* ' */ && marker !== 0x28 /* ( */) { return state }
+
+    start++
+    pos++
+
+    // if opening marker is "(", switch it to closing marker ")"
+    if (marker === 0x28) { marker = 0x29 }
+
+    state.marker = marker
+  }
+
+  while (pos < max) {
+    code = str.charCodeAt(pos)
+    if (code === state.marker) {
+      state.pos = pos + 1
+      state.str += unescapeAll(str.slice(start, pos))
+      state.ok = true
+      return state
+    } else if (code === 0x28 /* ( */ && state.marker === 0x29 /* ) */) {
+      return state
+    } else if (code === 0x5C /* \ */ && pos + 1 < max) {
+      pos++
+    }
+
+    pos++
+  }
+
+  // no closing marker found, but this link title may continue on the next line (for references)
+  state.can_continue = true
+  state.str += unescapeAll(str.slice(start, pos))
+  return state
+}

mcp-server/node_modules/markdown-it/lib/index.mjs

@@ -0,0 +1,565 @@
+// Main parser class
+
+import * as utils from './common/utils.mjs'
+import * as helpers from './helpers/index.mjs'
+import Renderer from './renderer.mjs'
+import ParserCore from './parser_core.mjs'
+import ParserBlock from './parser_block.mjs'
+import ParserInline from './parser_inline.mjs'
+import LinkifyIt from 'linkify-it'
+import * as mdurl from 'mdurl'
+import punycode from 'punycode.js'
+
+import cfg_default from './presets/default.mjs'
+import cfg_zero from './presets/zero.mjs'
+import cfg_commonmark from './presets/commonmark.mjs'
+
+const config = {
+  default: cfg_default,
+  zero: cfg_zero,
+  commonmark: cfg_commonmark
+}
+
+//
+// This validator can prohibit more than really needed to prevent XSS. It's a
+// tradeoff to keep code simple and to be secure by default.
+//
+// If you need different setup - override validator method as you wish. Or
+// replace it with dummy function and use external sanitizer.
+//
+
+const BAD_PROTO_RE = /^(vbscript|javascript|file|data):/
+const GOOD_DATA_RE = /^data:image\/(gif|png|jpeg|webp);/
+
+function validateLink (url) {
+  // url should be normalized at this point, and existing entities are decoded
+  const str = url.trim().toLowerCase()
+
+  return BAD_PROTO_RE.test(str) ? GOOD_DATA_RE.test(str) : true
+}
+
+const RECODE_HOSTNAME_FOR = ['http:', 'https:', 'mailto:']
+
+function normalizeLink (url) {
+  const parsed = mdurl.parse(url, true)
+
+  if (parsed.hostname) {
+    // Encode hostnames in urls like:
+    // `http://host/`, `https://host/`, `mailto:user@host`, `//host/`
+    //
+    // We don't encode unknown schemas, because it's likely that we encode
+    // something we shouldn't (e.g. `skype:name` treated as `skype:host`)
+    //
+    if (!parsed.protocol || RECODE_HOSTNAME_FOR.indexOf(parsed.protocol) >= 0) {
+      try {
+        parsed.hostname = punycode.toASCII(parsed.hostname)
+      } catch (er) { /**/ }
+    }
+  }
+
+  return mdurl.encode(mdurl.format(parsed))
+}
+
+function normalizeLinkText (url) {
+  const parsed = mdurl.parse(url, true)
+
+  if (parsed.hostname) {
+    // Encode hostnames in urls like:
+    // `http://host/`, `https://host/`, `mailto:user@host`, `//host/`
+    //
+    // We don't encode unknown schemas, because it's likely that we encode
+    // something we shouldn't (e.g. `skype:name` treated as `skype:host`)
+    //
+    if (!parsed.protocol || RECODE_HOSTNAME_FOR.indexOf(parsed.protocol) >= 0) {
+      try {
+        parsed.hostname = punycode.toUnicode(parsed.hostname)
+      } catch (er) { /**/ }
+    }
+  }
+
+  // add '%' to exclude list because of https://github.com/markdown-it/markdown-it/issues/720
+  return mdurl.decode(mdurl.format(parsed), mdurl.decode.defaultChars + '%')
+}
+
+/**
+ * class MarkdownIt
+ *
+ * Main parser/renderer class.
+ *
+ * ##### Usage
+ *
+ * ```javascript
+ * // node.js, "classic" way:
+ * var MarkdownIt = require('markdown-it'),
+ *     md = new MarkdownIt();
+ * var result = md.render('# markdown-it rulezz!');
+ *
+ * // node.js, the same, but with sugar:
+ * var md = require('markdown-it')();
+ * var result = md.render('# markdown-it rulezz!');
+ *
+ * // browser without AMD, added to "window" on script load
+ * // Note, there are no dash.
+ * var md = window.markdownit();
+ * var result = md.render('# markdown-it rulezz!');
+ * ```
+ *
+ * Single line rendering, without paragraph wrap:
+ *
+ * ```javascript
+ * var md = require('markdown-it')();
+ * var result = md.renderInline('__markdown-it__ rulezz!');
+ * ```
+ **/
+
+/**
+ * new MarkdownIt([presetName, options])
+ * - presetName (String): optional, `commonmark` / `zero`
+ * - options (Object)
+ *
+ * Creates parser instanse with given config. Can be called without `new`.
+ *
+ * ##### presetName
+ *
+ * MarkdownIt provides named presets as a convenience to quickly
+ * enable/disable active syntax rules and options for common use cases.
+ *
+ * - ["commonmark"](https://github.com/markdown-it/markdown-it/blob/master/lib/presets/commonmark.mjs) -
+ *   configures parser to strict [CommonMark](http://commonmark.org/) mode.
+ * - [default](https://github.com/markdown-it/markdown-it/blob/master/lib/presets/default.mjs) -
+ *   similar to GFM, used when no preset name given. Enables all available rules,
+ *   but still without html, typographer & autolinker.
+ * - ["zero"](https://github.com/markdown-it/markdown-it/blob/master/lib/presets/zero.mjs) -
+ *   all rules disabled. Useful to quickly setup your config via `.enable()`.
+ *   For example, when you need only `bold` and `italic` markup and nothing else.
+ *
+ * ##### options:
+ *
+ * - __html__ - `false`. Set `true` to enable HTML tags in source. Be careful!
+ *   That's not safe! You may need external sanitizer to protect output from XSS.
+ *   It's better to extend features via plugins, instead of enabling HTML.
+ * - __xhtmlOut__ - `false`. Set `true` to add '/' when closing single tags
+ *   (`<br />`). This is needed only for full CommonMark compatibility. In real
+ *   world you will need HTML output.
+ * - __breaks__ - `false`. Set `true` to convert `\n` in paragraphs into `<br>`.
+ * - __langPrefix__ - `language-`. CSS language class prefix for fenced blocks.
+ *   Can be useful for external highlighters.
+ * - __linkify__ - `false`. Set `true` to autoconvert URL-like text to links.
+ * - __typographer__  - `false`. Set `true` to enable [some language-neutral
+ *   replacement](https://github.com/markdown-it/markdown-it/blob/master/lib/rules_core/replacements.mjs) +
+ *   quotes beautification (smartquotes).
+ * - __quotes__ - `“”‘’`, String or Array. Double + single quotes replacement
+ *   pairs, when typographer enabled and smartquotes on. For example, you can
+ *   use `'«»„“'` for Russian, `'„“‚‘'` for German, and
+ *   `['«\xA0', '\xA0»', '‹\xA0', '\xA0›']` for French (including nbsp).
+ * - __highlight__ - `null`. Highlighter function for fenced code blocks.
+ *   Highlighter `function (str, lang)` should return escaped HTML. It can also
+ *   return empty string if the source was not changed and should be escaped
+ *   externaly. If result starts with <pre... internal wrapper is skipped.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * // commonmark mode
+ * var md = require('markdown-it')('commonmark');
+ *
+ * // default mode
+ * var md = require('markdown-it')();
+ *
+ * // enable everything
+ * var md = require('markdown-it')({
+ *   html: true,
+ *   linkify: true,
+ *   typographer: true
+ * });
+ * ```
+ *
+ * ##### Syntax highlighting
+ *
+ * ```js
+ * var hljs = require('highlight.js') // https://highlightjs.org/
+ *
+ * var md = require('markdown-it')({
+ *   highlight: function (str, lang) {
+ *     if (lang && hljs.getLanguage(lang)) {
+ *       try {
+ *         return hljs.highlight(str, { language: lang, ignoreIllegals: true }).value;
+ *       } catch (__) {}
+ *     }
+ *
+ *     return ''; // use external default escaping
+ *   }
+ * });
+ * ```
+ *
+ * Or with full wrapper override (if you need assign class to `<pre>` or `<code>`):
+ *
+ * ```javascript
+ * var hljs = require('highlight.js') // https://highlightjs.org/
+ *
+ * // Actual default values
+ * var md = require('markdown-it')({
+ *   highlight: function (str, lang) {
+ *     if (lang && hljs.getLanguage(lang)) {
+ *       try {
+ *         return '<pre><code class="hljs">' +
+ *                hljs.highlight(str, { language: lang, ignoreIllegals: true }).value +
+ *                '</code></pre>';
+ *       } catch (__) {}
+ *     }
+ *
+ *     return '<pre><code class="hljs">' + md.utils.escapeHtml(str) + '</code></pre>';
+ *   }
+ * });
+ * ```
+ *
+ **/
+function MarkdownIt (presetName, options) {
+  if (!(this instanceof MarkdownIt)) {
+    return new MarkdownIt(presetName, options)
+  }
+
+  if (!options) {
+    if (!utils.isString(presetName)) {
+      options = presetName || {}
+      presetName = 'default'
+    }
+  }
+
+  /**
+   * MarkdownIt#inline -> ParserInline
+   *
+   * Instance of [[ParserInline]]. You may need it to add new rules when
+   * writing plugins. For simple rules control use [[MarkdownIt.disable]] and
+   * [[MarkdownIt.enable]].
+   **/
+  this.inline = new ParserInline()
+
+  /**
+   * MarkdownIt#block -> ParserBlock
+   *
+   * Instance of [[ParserBlock]]. You may need it to add new rules when
+   * writing plugins. For simple rules control use [[MarkdownIt.disable]] and
+   * [[MarkdownIt.enable]].
+   **/
+  this.block = new ParserBlock()
+
+  /**
+   * MarkdownIt#core -> Core
+   *
+   * Instance of [[Core]] chain executor. You may need it to add new rules when
+   * writing plugins. For simple rules control use [[MarkdownIt.disable]] and
+   * [[MarkdownIt.enable]].
+   **/
+  this.core = new ParserCore()
+
+  /**
+   * MarkdownIt#renderer -> Renderer
+   *
+   * Instance of [[Renderer]]. Use it to modify output look. Or to add rendering
+   * rules for new token types, generated by plugins.
+   *
+   * ##### Example
+   *
+   * ```javascript
+   * var md = require('markdown-it')();
+   *
+   * function myToken(tokens, idx, options, env, self) {
+   *   //...
+   *   return result;
+   * };
+   *
+   * md.renderer.rules['my_token'] = myToken
+   * ```
+   *
+   * See [[Renderer]] docs and [source code](https://github.com/markdown-it/markdown-it/blob/master/lib/renderer.mjs).
+   **/
+  this.renderer = new Renderer()
+
+  /**
+   * MarkdownIt#linkify -> LinkifyIt
+   *
+   * [linkify-it](https://github.com/markdown-it/linkify-it) instance.
+   * Used by [linkify](https://github.com/markdown-it/markdown-it/blob/master/lib/rules_core/linkify.mjs)
+   * rule.
+   **/
+  this.linkify = new LinkifyIt()
+
+  /**
+   * MarkdownIt#validateLink(url) -> Boolean
+   *
+   * Link validation function. CommonMark allows too much in links. By default
+   * we disable `javascript:`, `vbscript:`, `file:` schemas, and almost all `data:...` schemas
+   * except some embedded image types.
+   *
+   * You can change this behaviour:
+   *
+   * ```javascript
+   * var md = require('markdown-it')();
+   * // enable everything
+   * md.validateLink = function () { return true; }
+   * ```
+   **/
+  this.validateLink = validateLink
+
+  /**
+   * MarkdownIt#normalizeLink(url) -> String
+   *
+   * Function used to encode link url to a machine-readable format,
+   * which includes url-encoding, punycode, etc.
+   **/
+  this.normalizeLink = normalizeLink
+
+  /**
+   * MarkdownIt#normalizeLinkText(url) -> String
+   *
+   * Function used to decode link url to a human-readable format`
+   **/
+  this.normalizeLinkText = normalizeLinkText
+
+  // Expose utils & helpers for easy acces from plugins
+
+  /**
+   * MarkdownIt#utils -> utils
+   *
+   * Assorted utility functions, useful to write plugins. See details
+   * [here](https://github.com/markdown-it/markdown-it/blob/master/lib/common/utils.mjs).
+   **/
+  this.utils = utils
+
+  /**
+   * MarkdownIt#helpers -> helpers
+   *
+   * Link components parser functions, useful to write plugins. See details
+   * [here](https://github.com/markdown-it/markdown-it/blob/master/lib/helpers).
+   **/
+  this.helpers = utils.assign({}, helpers)
+
+  this.options = {}
+  this.configure(presetName)
+
+  if (options) { this.set(options) }
+}
+
+/** chainable
+ * MarkdownIt.set(options)
+ *
+ * Set parser options (in the same format as in constructor). Probably, you
+ * will never need it, but you can change options after constructor call.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * var md = require('markdown-it')()
+ *             .set({ html: true, breaks: true })
+ *             .set({ typographer, true });
+ * ```
+ *
+ * __Note:__ To achieve the best possible performance, don't modify a
+ * `markdown-it` instance options on the fly. If you need multiple configurations
+ * it's best to create multiple instances and initialize each with separate
+ * config.
+ **/
+MarkdownIt.prototype.set = function (options) {
+  utils.assign(this.options, options)
+  return this
+}
+
+/** chainable, internal
+ * MarkdownIt.configure(presets)
+ *
+ * Batch load of all options and compenent settings. This is internal method,
+ * and you probably will not need it. But if you will - see available presets
+ * and data structure [here](https://github.com/markdown-it/markdown-it/tree/master/lib/presets)
+ *
+ * We strongly recommend to use presets instead of direct config loads. That
+ * will give better compatibility with next versions.
+ **/
+MarkdownIt.prototype.configure = function (presets) {
+  const self = this
+
+  if (utils.isString(presets)) {
+    const presetName = presets
+    presets = config[presetName]
+    if (!presets) { throw new Error('Wrong `markdown-it` preset "' + presetName + '", check name') }
+  }
+
+  if (!presets) { throw new Error('Wrong `markdown-it` preset, can\'t be empty') }
+
+  if (presets.options) { self.set(presets.options) }
+
+  if (presets.components) {
+    Object.keys(presets.components).forEach(function (name) {
+      if (presets.components[name].rules) {
+        self[name].ruler.enableOnly(presets.components[name].rules)
+      }
+      if (presets.components[name].rules2) {
+        self[name].ruler2.enableOnly(presets.components[name].rules2)
+      }
+    })
+  }
+  return this
+}
+
+/** chainable
+ * MarkdownIt.enable(list, ignoreInvalid)
+ * - list (String|Array): rule name or list of rule names to enable
+ * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found.
+ *
+ * Enable list or rules. It will automatically find appropriate components,
+ * containing rules with given names. If rule not found, and `ignoreInvalid`
+ * not set - throws exception.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * var md = require('markdown-it')()
+ *             .enable(['sub', 'sup'])
+ *             .disable('smartquotes');
+ * ```
+ **/
+MarkdownIt.prototype.enable = function (list, ignoreInvalid) {
+  let result = []
+
+  if (!Array.isArray(list)) { list = [list] }
+
+  ['core', 'block', 'inline'].forEach(function (chain) {
+    result = result.concat(this[chain].ruler.enable(list, true))
+  }, this)
+
+  result = result.concat(this.inline.ruler2.enable(list, true))
+
+  const missed = list.filter(function (name) { return result.indexOf(name) < 0 })
+
+  if (missed.length && !ignoreInvalid) {
+    throw new Error('MarkdownIt. Failed to enable unknown rule(s): ' + missed)
+  }
+
+  return this
+}
+
+/** chainable
+ * MarkdownIt.disable(list, ignoreInvalid)
+ * - list (String|Array): rule name or list of rule names to disable.
+ * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found.
+ *
+ * The same as [[MarkdownIt.enable]], but turn specified rules off.
+ **/
+MarkdownIt.prototype.disable = function (list, ignoreInvalid) {
+  let result = []
+
+  if (!Array.isArray(list)) { list = [list] }
+
+  ['core', 'block', 'inline'].forEach(function (chain) {
+    result = result.concat(this[chain].ruler.disable(list, true))
+  }, this)
+
+  result = result.concat(this.inline.ruler2.disable(list, true))
+
+  const missed = list.filter(function (name) { return result.indexOf(name) < 0 })
+
+  if (missed.length && !ignoreInvalid) {
+    throw new Error('MarkdownIt. Failed to disable unknown rule(s): ' + missed)
+  }
+  return this
+}
+
+/** chainable
+ * MarkdownIt.use(plugin, params)
+ *
+ * Load specified plugin with given params into current parser instance.
+ * It's just a sugar to call `plugin(md, params)` with curring.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * var iterator = require('markdown-it-for-inline');
+ * var md = require('markdown-it')()
+ *             .use(iterator, 'foo_replace', 'text', function (tokens, idx) {
+ *               tokens[idx].content = tokens[idx].content.replace(/foo/g, 'bar');
+ *             });
+ * ```
+ **/
+MarkdownIt.prototype.use = function (plugin /*, params, ... */) {
+  const args = [this].concat(Array.prototype.slice.call(arguments, 1))
+  plugin.apply(plugin, args)
+  return this
+}
+
+/** internal
+ * MarkdownIt.parse(src, env) -> Array
+ * - src (String): source string
+ * - env (Object): environment sandbox
+ *
+ * Parse input string and return list of block tokens (special token type
+ * "inline" will contain list of inline tokens). You should not call this
+ * method directly, until you write custom renderer (for example, to produce
+ * AST).
+ *
+ * `env` is used to pass data between "distributed" rules and return additional
+ * metadata like reference info, needed for the renderer. It also can be used to
+ * inject data in specific cases. Usually, you will be ok to pass `{}`,
+ * and then pass updated object to renderer.
+ **/
+MarkdownIt.prototype.parse = function (src, env) {
+  if (typeof src !== 'string') {
+    throw new Error('Input data should be a String')
+  }
+
+  const state = new this.core.State(src, this, env)
+
+  this.core.process(state)
+
+  return state.tokens
+}
+
+/**
+ * MarkdownIt.render(src [, env]) -> String
+ * - src (String): source string
+ * - env (Object): environment sandbox
+ *
+ * Render markdown string into html. It does all magic for you :).
+ *
+ * `env` can be used to inject additional metadata (`{}` by default).
+ * But you will not need it with high probability. See also comment
+ * in [[MarkdownIt.parse]].
+ **/
+MarkdownIt.prototype.render = function (src, env) {
+  env = env || {}
+
+  return this.renderer.render(this.parse(src, env), this.options, env)
+}
+
+/** internal
+ * MarkdownIt.parseInline(src, env) -> Array
+ * - src (String): source string
+ * - env (Object): environment sandbox
+ *
+ * The same as [[MarkdownIt.parse]] but skip all block rules. It returns the
+ * block tokens list with the single `inline` element, containing parsed inline
+ * tokens in `children` property. Also updates `env` object.
+ **/
+MarkdownIt.prototype.parseInline = function (src, env) {
+  const state = new this.core.State(src, this, env)
+
+  state.inlineMode = true
+  this.core.process(state)
+
+  return state.tokens
+}
+
+/**
+ * MarkdownIt.renderInline(src [, env]) -> String
+ * - src (String): source string
+ * - env (Object): environment sandbox
+ *
+ * Similar to [[MarkdownIt.render]] but for single paragraph content. Result
+ * will NOT be wrapped into `<p>` tags.
+ **/
+MarkdownIt.prototype.renderInline = function (src, env) {
+  env = env || {}
+
+  return this.renderer.render(this.parseInline(src, env), this.options, env)
+}
+
+export default MarkdownIt

mcp-server/node_modules/markdown-it/lib/parser_block.mjs

@@ -0,0 +1,134 @@
+/** internal
+ * class ParserBlock
+ *
+ * Block-level tokenizer.
+ **/
+
+import Ruler from './ruler.mjs'
+import StateBlock from './rules_block/state_block.mjs'
+
+import r_table from './rules_block/table.mjs'
+import r_code from './rules_block/code.mjs'
+import r_fence from './rules_block/fence.mjs'
+import r_blockquote from './rules_block/blockquote.mjs'
+import r_hr from './rules_block/hr.mjs'
+import r_list from './rules_block/list.mjs'
+import r_reference from './rules_block/reference.mjs'
+import r_html_block from './rules_block/html_block.mjs'
+import r_heading from './rules_block/heading.mjs'
+import r_lheading from './rules_block/lheading.mjs'
+import r_paragraph from './rules_block/paragraph.mjs'
+
+const _rules = [
+  // First 2 params - rule name & source. Secondary array - list of rules,
+  // which can be terminated by this one.
+  ['table',      r_table,      ['paragraph', 'reference']],
+  ['code',       r_code],
+  ['fence',      r_fence,      ['paragraph', 'reference', 'blockquote', 'list']],
+  ['blockquote', r_blockquote, ['paragraph', 'reference', 'blockquote', 'list']],
+  ['hr',         r_hr,         ['paragraph', 'reference', 'blockquote', 'list']],
+  ['list',       r_list,       ['paragraph', 'reference', 'blockquote']],
+  ['reference',  r_reference],
+  ['html_block', r_html_block, ['paragraph', 'reference', 'blockquote']],
+  ['heading',    r_heading,    ['paragraph', 'reference', 'blockquote']],
+  ['lheading',   r_lheading],
+  ['paragraph',  r_paragraph]
+]
+
+/**
+ * new ParserBlock()
+ **/
+function ParserBlock () {
+  /**
+   * ParserBlock#ruler -> Ruler
+   *
+   * [[Ruler]] instance. Keep configuration of block rules.
+   **/
+  this.ruler = new Ruler()
+
+  for (let i = 0; i < _rules.length; i++) {
+    this.ruler.push(_rules[i][0], _rules[i][1], { alt: (_rules[i][2] || []).slice() })
+  }
+}
+
+// Generate tokens for input range
+//
+ParserBlock.prototype.tokenize = function (state, startLine, endLine) {
+  const rules = this.ruler.getRules('')
+  const len = rules.length
+  const maxNesting = state.md.options.maxNesting
+  let line = startLine
+  let hasEmptyLines = false
+
+  while (line < endLine) {
+    state.line = line = state.skipEmptyLines(line)
+    if (line >= endLine) { break }
+
+    // Termination condition for nested calls.
+    // Nested calls currently used for blockquotes & lists
+    if (state.sCount[line] < state.blkIndent) { break }
+
+    // If nesting level exceeded - skip tail to the end. That's not ordinary
+    // situation and we should not care about content.
+    if (state.level >= maxNesting) {
+      state.line = endLine
+      break
+    }
+
+    // Try all possible rules.
+    // On success, rule should:
+    //
+    // - update `state.line`
+    // - update `state.tokens`
+    // - return true
+    const prevLine = state.line
+    let ok = false
+
+    for (let i = 0; i < len; i++) {
+      ok = rules[i](state, line, endLine, false)
+      if (ok) {
+        if (prevLine >= state.line) {
+          throw new Error("block rule didn't increment state.line")
+        }
+        break
+      }
+    }
+
+    // this can only happen if user disables paragraph rule
+    if (!ok) throw new Error('none of the block rules matched')
+
+    // set state.tight if we had an empty line before current tag
+    // i.e. latest empty line should not count
+    state.tight = !hasEmptyLines
+
+    // paragraph might "eat" one newline after it in nested lists
+    if (state.isEmpty(state.line - 1)) {
+      hasEmptyLines = true
+    }
+
+    line = state.line
+
+    if (line < endLine && state.isEmpty(line)) {
+      hasEmptyLines = true
+      line++
+      state.line = line
+    }
+  }
+}
+
+/**
+ * ParserBlock.parse(str, md, env, outTokens)
+ *
+ * Process input string and push block tokens into `outTokens`
+ **/
+ParserBlock.prototype.parse = function (src, md, env, outTokens) {
+  if (!src) { return }
+
+  const state = new this.State(src, md, env, outTokens)
+
+  this.tokenize(state, state.line, state.lineMax)
+}
+
+ParserBlock.prototype.State = StateBlock
+
+export default ParserBlock

mcp-server/node_modules/markdown-it/lib/parser_core.mjs

@@ -0,0 +1,62 @@
+/** internal
+ * class Core
+ *
+ * Top-level rules executor. Glues block/inline parsers and does intermediate
+ * transformations.
+ **/
+
+import Ruler from './ruler.mjs'
+import StateCore from './rules_core/state_core.mjs'
+
+import r_normalize from './rules_core/normalize.mjs'
+import r_block from './rules_core/block.mjs'
+import r_inline from './rules_core/inline.mjs'
+import r_linkify from './rules_core/linkify.mjs'
+import r_replacements from './rules_core/replacements.mjs'
+import r_smartquotes from './rules_core/smartquotes.mjs'
+import r_text_join from './rules_core/text_join.mjs'
+
+const _rules = [
+  ['normalize',      r_normalize],
+  ['block',          r_block],
+  ['inline',         r_inline],
+  ['linkify',        r_linkify],
+  ['replacements',   r_replacements],
+  ['smartquotes',    r_smartquotes],
+  // `text_join` finds `text_special` tokens (for escape sequences)
+  // and joins them with the rest of the text
+  ['text_join',      r_text_join]
+]
+
+/**
+ * new Core()
+ **/
+function Core () {
+  /**
+   * Core#ruler -> Ruler
+   *
+   * [[Ruler]] instance. Keep configuration of core rules.
+   **/
+  this.ruler = new Ruler()
+
+  for (let i = 0; i < _rules.length; i++) {
+    this.ruler.push(_rules[i][0], _rules[i][1])
+  }
+}
+
+/**
+ * Core.process(state)
+ *
+ * Executes core chain rules.
+ **/
+Core.prototype.process = function (state) {
+  const rules = this.ruler.getRules('')
+
+  for (let i = 0, l = rules.length; i < l; i++) {
+    rules[i](state)
+  }
+}
+
+Core.prototype.State = StateCore
+
+export default Core

mcp-server/node_modules/markdown-it/lib/parser_inline.mjs

@@ -0,0 +1,197 @@
+/** internal
+ * class ParserInline
+ *
+ * Tokenizes paragraph content.
+ **/
+
+import Ruler from './ruler.mjs'
+import StateInline from './rules_inline/state_inline.mjs'
+
+import r_text from './rules_inline/text.mjs'
+import r_linkify from './rules_inline/linkify.mjs'
+import r_newline from './rules_inline/newline.mjs'
+import r_escape from './rules_inline/escape.mjs'
+import r_backticks from './rules_inline/backticks.mjs'
+import r_strikethrough from './rules_inline/strikethrough.mjs'
+import r_emphasis from './rules_inline/emphasis.mjs'
+import r_link from './rules_inline/link.mjs'
+import r_image from './rules_inline/image.mjs'
+import r_autolink from './rules_inline/autolink.mjs'
+import r_html_inline from './rules_inline/html_inline.mjs'
+import r_entity from './rules_inline/entity.mjs'
+
+import r_balance_pairs from './rules_inline/balance_pairs.mjs'
+import r_fragments_join from './rules_inline/fragments_join.mjs'
+
+// Parser rules
+
+const _rules = [
+  ['text',            r_text],
+  ['linkify',         r_linkify],
+  ['newline',         r_newline],
+  ['escape',          r_escape],
+  ['backticks',       r_backticks],
+  ['strikethrough',   r_strikethrough.tokenize],
+  ['emphasis',        r_emphasis.tokenize],
+  ['link',            r_link],
+  ['image',           r_image],
+  ['autolink',        r_autolink],
+  ['html_inline',     r_html_inline],
+  ['entity',          r_entity]
+]
+
+// `rule2` ruleset was created specifically for emphasis/strikethrough
+// post-processing and may be changed in the future.
+//
+// Don't use this for anything except pairs (plugins working with `balance_pairs`).
+//
+const _rules2 = [
+  ['balance_pairs',   r_balance_pairs],
+  ['strikethrough',   r_strikethrough.postProcess],
+  ['emphasis',        r_emphasis.postProcess],
+  // rules for pairs separate '**' into its own text tokens, which may be left unused,
+  // rule below merges unused segments back with the rest of the text
+  ['fragments_join',  r_fragments_join]
+]
+
+/**
+ * new ParserInline()
+ **/
+function ParserInline () {
+  /**
+   * ParserInline#ruler -> Ruler
+   *
+   * [[Ruler]] instance. Keep configuration of inline rules.
+   **/
+  this.ruler = new Ruler()
+
+  for (let i = 0; i < _rules.length; i++) {
+    this.ruler.push(_rules[i][0], _rules[i][1])
+  }
+
+  /**
+   * ParserInline#ruler2 -> Ruler
+   *
+   * [[Ruler]] instance. Second ruler used for post-processing
+   * (e.g. in emphasis-like rules).
+   **/
+  this.ruler2 = new Ruler()
+
+  for (let i = 0; i < _rules2.length; i++) {
+    this.ruler2.push(_rules2[i][0], _rules2[i][1])
+  }
+}
+
+// Skip single token by running all rules in validation mode;
+// returns `true` if any rule reported success
+//
+ParserInline.prototype.skipToken = function (state) {
+  const pos = state.pos
+  const rules = this.ruler.getRules('')
+  const len = rules.length
+  const maxNesting = state.md.options.maxNesting
+  const cache = state.cache
+
+  if (typeof cache[pos] !== 'undefined') {
+    state.pos = cache[pos]
+    return
+  }
+
+  let ok = false
+
+  if (state.level < maxNesting) {
+    for (let i = 0; i < len; i++) {
+      // Increment state.level and decrement it later to limit recursion.
+      // It's harmless to do here, because no tokens are created. But ideally,
+      // we'd need a separate private state variable for this purpose.
+      //
+      state.level++
+      ok = rules[i](state, true)
+      state.level--
+
+      if (ok) {
+        if (pos >= state.pos) { throw new Error("inline rule didn't increment state.pos") }
+        break
+      }
+    }
+  } else {
+    // Too much nesting, just skip until the end of the paragraph.
+    //
+    // NOTE: this will cause links to behave incorrectly in the following case,
+    //       when an amount of `[` is exactly equal to `maxNesting + 1`:
+    //
+    //       [[[[[[[[[[[[[[[[[[[[[foo]()
+    //
+    // TODO: remove this workaround when CM standard will allow nested links
+    //       (we can replace it by preventing links from being parsed in
+    //       validation mode)
+    //
+    state.pos = state.posMax
+  }
+
+  if (!ok) { state.pos++ }
+  cache[pos] = state.pos
+}
+
+// Generate tokens for input range
+//
+ParserInline.prototype.tokenize = function (state) {
+  const rules = this.ruler.getRules('')
+  const len = rules.length
+  const end = state.posMax
+  const maxNesting = state.md.options.maxNesting
+
+  while (state.pos < end) {
+    // Try all possible rules.
+    // On success, rule should:
+    //
+    // - update `state.pos`
+    // - update `state.tokens`
+    // - return true
+    const prevPos = state.pos
+    let ok = false
+
+    if (state.level < maxNesting) {
+      for (let i = 0; i < len; i++) {
+        ok = rules[i](state, false)
+        if (ok) {
+          if (prevPos >= state.pos) { throw new Error("inline rule didn't increment state.pos") }
+          break
+        }
+      }
+    }
+
+    if (ok) {
+      if (state.pos >= end) { break }
+      continue
+    }
+
+    state.pending += state.src[state.pos++]
+  }
+
+  if (state.pending) {
+    state.pushPending()
+  }
+}
+
+/**
+ * ParserInline.parse(str, md, env, outTokens)
+ *
+ * Process input string and push inline tokens into `outTokens`
+ **/
+ParserInline.prototype.parse = function (str, md, env, outTokens) {
+  const state = new this.State(str, md, env, outTokens)
+
+  this.tokenize(state)
+
+  const rules = this.ruler2.getRules('')
+  const len = rules.length
+
+  for (let i = 0; i < len; i++) {
+    rules[i](state)
+  }
+}
+
+ParserInline.prototype.State = StateInline
+
+export default ParserInline

mcp-server/node_modules/markdown-it/lib/presets/commonmark.mjs

@@ -0,0 +1,88 @@
+// Commonmark default options
+
+export default {
+  options: {
+    // Enable HTML tags in source
+    html: true,
+
+    // Use '/' to close single tags (<br />)
+    xhtmlOut: true,
+
+    // Convert '\n' in paragraphs into <br>
+    breaks: false,
+
+    // CSS language prefix for fenced blocks
+    langPrefix: 'language-',
+
+    // autoconvert URL-like texts to links
+    linkify: false,
+
+    // Enable some language-neutral replacements + quotes beautification
+    typographer: false,
+
+    // Double + single quotes replacement pairs, when typographer enabled,
+    // and smartquotes on. Could be either a String or an Array.
+    //
+    // For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+    // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+    quotes: '\u201c\u201d\u2018\u2019', /* “”‘’ */
+
+    // Highlighter function. Should return escaped HTML,
+    // or '' if the source string is not changed and should be escaped externaly.
+    // If result starts with <pre... internal wrapper is skipped.
+    //
+    // function (/*str, lang*/) { return ''; }
+    //
+    highlight: null,
+
+    // Internal protection, recursion limit
+    maxNesting: 20
+  },
+
+  components: {
+
+    core: {
+      rules: [
+        'normalize',
+        'block',
+        'inline',
+        'text_join'
+      ]
+    },
+
+    block: {
+      rules: [
+        'blockquote',
+        'code',
+        'fence',
+        'heading',
+        'hr',
+        'html_block',
+        'lheading',
+        'list',
+        'reference',
+        'paragraph'
+      ]
+    },
+
+    inline: {
+      rules: [
+        'autolink',
+        'backticks',
+        'emphasis',
+        'entity',
+        'escape',
+        'html_inline',
+        'image',
+        'link',
+        'newline',
+        'text'
+      ],
+      rules2: [
+        'balance_pairs',
+        'emphasis',
+        'fragments_join'
+      ]
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/presets/default.mjs

@@ -0,0 +1,47 @@
+// markdown-it default options
+
+export default {
+  options: {
+    // Enable HTML tags in source
+    html: false,
+
+    // Use '/' to close single tags (<br />)
+    xhtmlOut: false,
+
+    // Convert '\n' in paragraphs into <br>
+    breaks: false,
+
+    // CSS language prefix for fenced blocks
+    langPrefix: 'language-',
+
+    // autoconvert URL-like texts to links
+    linkify: false,
+
+    // Enable some language-neutral replacements + quotes beautification
+    typographer: false,
+
+    // Double + single quotes replacement pairs, when typographer enabled,
+    // and smartquotes on. Could be either a String or an Array.
+    //
+    // For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+    // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+    quotes: '\u201c\u201d\u2018\u2019', /* “”‘’ */
+
+    // Highlighter function. Should return escaped HTML,
+    // or '' if the source string is not changed and should be escaped externaly.
+    // If result starts with <pre... internal wrapper is skipped.
+    //
+    // function (/*str, lang*/) { return ''; }
+    //
+    highlight: null,
+
+    // Internal protection, recursion limit
+    maxNesting: 100
+  },
+
+  components: {
+    core: {},
+    block: {},
+    inline: {}
+  }
+}

mcp-server/node_modules/markdown-it/lib/presets/zero.mjs

@@ -0,0 +1,70 @@
+// "Zero" preset, with nothing enabled. Useful for manual configuring of simple
+// modes. For example, to parse bold/italic only.
+
+export default {
+  options: {
+    // Enable HTML tags in source
+    html: false,
+
+    // Use '/' to close single tags (<br />)
+    xhtmlOut: false,
+
+    // Convert '\n' in paragraphs into <br>
+    breaks: false,
+
+    // CSS language prefix for fenced blocks
+    langPrefix: 'language-',
+
+    // autoconvert URL-like texts to links
+    linkify: false,
+
+    // Enable some language-neutral replacements + quotes beautification
+    typographer: false,
+
+    // Double + single quotes replacement pairs, when typographer enabled,
+    // and smartquotes on. Could be either a String or an Array.
+    //
+    // For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+    // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+    quotes: '\u201c\u201d\u2018\u2019', /* “”‘’ */
+
+    // Highlighter function. Should return escaped HTML,
+    // or '' if the source string is not changed and should be escaped externaly.
+    // If result starts with <pre... internal wrapper is skipped.
+    //
+    // function (/*str, lang*/) { return ''; }
+    //
+    highlight: null,
+
+    // Internal protection, recursion limit
+    maxNesting: 20
+  },
+
+  components: {
+
+    core: {
+      rules: [
+        'normalize',
+        'block',
+        'inline',
+        'text_join'
+      ]
+    },
+
+    block: {
+      rules: [
+        'paragraph'
+      ]
+    },
+
+    inline: {
+      rules: [
+        'text'
+      ],
+      rules2: [
+        'balance_pairs',
+        'fragments_join'
+      ]
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/renderer.mjs

@@ -0,0 +1,322 @@
+/**
+ * class Renderer
+ *
+ * Generates HTML from parsed token stream. Each instance has independent
+ * copy of rules. Those can be rewritten with ease. Also, you can add new
+ * rules if you create plugin and adds new token types.
+ **/
+
+import { assign, unescapeAll, escapeHtml } from './common/utils.mjs'
+
+const default_rules = {}
+
+default_rules.code_inline = function (tokens, idx, options, env, slf) {
+  const token = tokens[idx]
+
+  return  '<code' + slf.renderAttrs(token) + '>' +
+          escapeHtml(token.content) +
+          '</code>'
+}
+
+default_rules.code_block = function (tokens, idx, options, env, slf) {
+  const token = tokens[idx]
+
+  return  '<pre' + slf.renderAttrs(token) + '><code>' +
+          escapeHtml(tokens[idx].content) +
+          '</code></pre>\n'
+}
+
+default_rules.fence = function (tokens, idx, options, env, slf) {
+  const token = tokens[idx]
+  const info = token.info ? unescapeAll(token.info).trim() : ''
+  let langName = ''
+  let langAttrs = ''
+
+  if (info) {
+    const arr = info.split(/(\s+)/g)
+    langName = arr[0]
+    langAttrs = arr.slice(2).join('')
+  }
+
+  let highlighted
+  if (options.highlight) {
+    highlighted = options.highlight(token.content, langName, langAttrs) || escapeHtml(token.content)
+  } else {
+    highlighted = escapeHtml(token.content)
+  }
+
+  if (highlighted.indexOf('<pre') === 0) {
+    return highlighted + '\n'
+  }
+
+  // If language exists, inject class gently, without modifying original token.
+  // May be, one day we will add .deepClone() for token and simplify this part, but
+  // now we prefer to keep things local.
+  if (info) {
+    const i = token.attrIndex('class')
+    const tmpAttrs = token.attrs ? token.attrs.slice() : []
+
+    if (i < 0) {
+      tmpAttrs.push(['class', options.langPrefix + langName])
+    } else {
+      tmpAttrs[i] = tmpAttrs[i].slice()
+      tmpAttrs[i][1] += ' ' + options.langPrefix + langName
+    }
+
+    // Fake token just to render attributes
+    const tmpToken = {
+      attrs: tmpAttrs
+    }
+
+    return `<pre><code${slf.renderAttrs(tmpToken)}>${highlighted}</code></pre>\n`
+  }
+
+  return `<pre><code${slf.renderAttrs(token)}>${highlighted}</code></pre>\n`
+}
+
+default_rules.image = function (tokens, idx, options, env, slf) {
+  const token = tokens[idx]
+
+  // "alt" attr MUST be set, even if empty. Because it's mandatory and
+  // should be placed on proper position for tests.
+  //
+  // Replace content with actual value
+
+  token.attrs[token.attrIndex('alt')][1] =
+    slf.renderInlineAsText(token.children, options, env)
+
+  return slf.renderToken(tokens, idx, options)
+}
+
+default_rules.hardbreak = function (tokens, idx, options /*, env */) {
+  return options.xhtmlOut ? '<br />\n' : '<br>\n'
+}
+default_rules.softbreak = function (tokens, idx, options /*, env */) {
+  return options.breaks ? (options.xhtmlOut ? '<br />\n' : '<br>\n') : '\n'
+}
+
+default_rules.text = function (tokens, idx /*, options, env */) {
+  return escapeHtml(tokens[idx].content)
+}
+
+default_rules.html_block = function (tokens, idx /*, options, env */) {
+  return tokens[idx].content
+}
+default_rules.html_inline = function (tokens, idx /*, options, env */) {
+  return tokens[idx].content
+}
+
+/**
+ * new Renderer()
+ *
+ * Creates new [[Renderer]] instance and fill [[Renderer#rules]] with defaults.
+ **/
+function Renderer () {
+  /**
+   * Renderer#rules -> Object
+   *
+   * Contains render rules for tokens. Can be updated and extended.
+   *
+   * ##### Example
+   *
+   * ```javascript
+   * var md = require('markdown-it')();
+   *
+   * md.renderer.rules.strong_open  = function () { return '<b>'; };
+   * md.renderer.rules.strong_close = function () { return '</b>'; };
+   *
+   * var result = md.renderInline(...);
+   * ```
+   *
+   * Each rule is called as independent static function with fixed signature:
+   *
+   * ```javascript
+   * function my_token_render(tokens, idx, options, env, renderer) {
+   *   // ...
+   *   return renderedHTML;
+   * }
+   * ```
+   *
+   * See [source code](https://github.com/markdown-it/markdown-it/blob/master/lib/renderer.mjs)
+   * for more details and examples.
+   **/
+  this.rules = assign({}, default_rules)
+}
+
+/**
+ * Renderer.renderAttrs(token) -> String
+ *
+ * Render token attributes to string.
+ **/
+Renderer.prototype.renderAttrs = function renderAttrs (token) {
+  let i, l, result
+
+  if (!token.attrs) { return '' }
+
+  result = ''
+
+  for (i = 0, l = token.attrs.length; i < l; i++) {
+    result += ' ' + escapeHtml(token.attrs[i][0]) + '="' + escapeHtml(token.attrs[i][1]) + '"'
+  }
+
+  return result
+}
+
+/**
+ * Renderer.renderToken(tokens, idx, options) -> String
+ * - tokens (Array): list of tokens
+ * - idx (Numbed): token index to render
+ * - options (Object): params of parser instance
+ *
+ * Default token renderer. Can be overriden by custom function
+ * in [[Renderer#rules]].
+ **/
+Renderer.prototype.renderToken = function renderToken (tokens, idx, options) {
+  const token = tokens[idx]
+  let result = ''
+
+  // Tight list paragraphs
+  if (token.hidden) {
+    return ''
+  }
+
+  // Insert a newline between hidden paragraph and subsequent opening
+  // block-level tag.
+  //
+  // For example, here we should insert a newline before blockquote:
+  //  - a
+  //    >
+  //
+  if (token.block && token.nesting !== -1 && idx && tokens[idx - 1].hidden) {
+    result += '\n'
+  }
+
+  // Add token name, e.g. `<img`
+  result += (token.nesting === -1 ? '</' : '<') + token.tag
+
+  // Encode attributes, e.g. `<img src="foo"`
+  result += this.renderAttrs(token)
+
+  // Add a slash for self-closing tags, e.g. `<img src="foo" /`
+  if (token.nesting === 0 && options.xhtmlOut) {
+    result += ' /'
+  }
+
+  // Check if we need to add a newline after this tag
+  let needLf = false
+  if (token.block) {
+    needLf = true
+
+    if (token.nesting === 1) {
+      if (idx + 1 < tokens.length) {
+        const nextToken = tokens[idx + 1]
+
+        if (nextToken.type === 'inline' || nextToken.hidden) {
+          // Block-level tag containing an inline tag.
+          //
+          needLf = false
+        } else if (nextToken.nesting === -1 && nextToken.tag === token.tag) {
+          // Opening tag + closing tag of the same type. E.g. `<li></li>`.
+          //
+          needLf = false
+        }
+      }
+    }
+  }
+
+  result += needLf ? '>\n' : '>'
+
+  return result
+}
+
+/**
+ * Renderer.renderInline(tokens, options, env) -> String
+ * - tokens (Array): list on block tokens to render
+ * - options (Object): params of parser instance
+ * - env (Object): additional data from parsed input (references, for example)
+ *
+ * The same as [[Renderer.render]], but for single token of `inline` type.
+ **/
+Renderer.prototype.renderInline = function (tokens, options, env) {
+  let result = ''
+  const rules = this.rules
+
+  for (let i = 0, len = tokens.length; i < len; i++) {
+    const type = tokens[i].type
+
+    if (typeof rules[type] !== 'undefined') {
+      result += rules[type](tokens, i, options, env, this)
+    } else {
+      result += this.renderToken(tokens, i, options)
+    }
+  }
+
+  return result
+}
+
+/** internal
+ * Renderer.renderInlineAsText(tokens, options, env) -> String
+ * - tokens (Array): list on block tokens to render
+ * - options (Object): params of parser instance
+ * - env (Object): additional data from parsed input (references, for example)
+ *
+ * Special kludge for image `alt` attributes to conform CommonMark spec.
+ * Don't try to use it! Spec requires to show `alt` content with stripped markup,
+ * instead of simple escaping.
+ **/
+Renderer.prototype.renderInlineAsText = function (tokens, options, env) {
+  let result = ''
+
+  for (let i = 0, len = tokens.length; i < len; i++) {
+    switch (tokens[i].type) {
+      case 'text':
+        result += tokens[i].content
+        break
+      case 'image':
+        result += this.renderInlineAsText(tokens[i].children, options, env)
+        break
+      case 'html_inline':
+      case 'html_block':
+        result += tokens[i].content
+        break
+      case 'softbreak':
+      case 'hardbreak':
+        result += '\n'
+        break
+      default:
+        // all other tokens are skipped
+    }
+  }
+
+  return result
+}
+
+/**
+ * Renderer.render(tokens, options, env) -> String
+ * - tokens (Array): list on block tokens to render
+ * - options (Object): params of parser instance
+ * - env (Object): additional data from parsed input (references, for example)
+ *
+ * Takes token stream and generates HTML. Probably, you will never need to call
+ * this method directly.
+ **/
+Renderer.prototype.render = function (tokens, options, env) {
+  let result = ''
+  const rules = this.rules
+
+  for (let i = 0, len = tokens.length; i < len; i++) {
+    const type = tokens[i].type
+
+    if (type === 'inline') {
+      result += this.renderInline(tokens[i].children, options, env)
+    } else if (typeof rules[type] !== 'undefined') {
+      result += rules[type](tokens, i, options, env, this)
+    } else {
+      result += this.renderToken(tokens, i, options, env)
+    }
+  }
+
+  return result
+}
+
+export default Renderer

mcp-server/node_modules/markdown-it/lib/ruler.mjs

@@ -0,0 +1,340 @@
+/**
+ * class Ruler
+ *
+ * Helper class, used by [[MarkdownIt#core]], [[MarkdownIt#block]] and
+ * [[MarkdownIt#inline]] to manage sequences of functions (rules):
+ *
+ * - keep rules in defined order
+ * - assign the name to each rule
+ * - enable/disable rules
+ * - add/replace rules
+ * - allow assign rules to additional named chains (in the same)
+ * - cacheing lists of active rules
+ *
+ * You will not need use this class directly until write plugins. For simple
+ * rules control use [[MarkdownIt.disable]], [[MarkdownIt.enable]] and
+ * [[MarkdownIt.use]].
+ **/
+
+/**
+ * new Ruler()
+ **/
+function Ruler () {
+  // List of added rules. Each element is:
+  //
+  // {
+  //   name: XXX,
+  //   enabled: Boolean,
+  //   fn: Function(),
+  //   alt: [ name2, name3 ]
+  // }
+  //
+  this.__rules__ = []
+
+  // Cached rule chains.
+  //
+  // First level - chain name, '' for default.
+  // Second level - diginal anchor for fast filtering by charcodes.
+  //
+  this.__cache__ = null
+}
+
+// Helper methods, should not be used directly
+
+// Find rule index by name
+//
+Ruler.prototype.__find__ = function (name) {
+  for (let i = 0; i < this.__rules__.length; i++) {
+    if (this.__rules__[i].name === name) {
+      return i
+    }
+  }
+  return -1
+}
+
+// Build rules lookup cache
+//
+Ruler.prototype.__compile__ = function () {
+  const self = this
+  const chains = ['']
+
+  // collect unique names
+  self.__rules__.forEach(function (rule) {
+    if (!rule.enabled) { return }
+
+    rule.alt.forEach(function (altName) {
+      if (chains.indexOf(altName) < 0) {
+        chains.push(altName)
+      }
+    })
+  })
+
+  self.__cache__ = {}
+
+  chains.forEach(function (chain) {
+    self.__cache__[chain] = []
+    self.__rules__.forEach(function (rule) {
+      if (!rule.enabled) { return }
+
+      if (chain && rule.alt.indexOf(chain) < 0) { return }
+
+      self.__cache__[chain].push(rule.fn)
+    })
+  })
+}
+
+/**
+ * Ruler.at(name, fn [, options])
+ * - name (String): rule name to replace.
+ * - fn (Function): new rule function.
+ * - options (Object): new rule options (not mandatory).
+ *
+ * Replace rule by name with new function & options. Throws error if name not
+ * found.
+ *
+ * ##### Options:
+ *
+ * - __alt__ - array with names of "alternate" chains.
+ *
+ * ##### Example
+ *
+ * Replace existing typographer replacement rule with new one:
+ *
+ * ```javascript
+ * var md = require('markdown-it')();
+ *
+ * md.core.ruler.at('replacements', function replace(state) {
+ *   //...
+ * });
+ * ```
+ **/
+Ruler.prototype.at = function (name, fn, options) {
+  const index = this.__find__(name)
+  const opt = options || {}
+
+  if (index === -1) { throw new Error('Parser rule not found: ' + name) }
+
+  this.__rules__[index].fn = fn
+  this.__rules__[index].alt = opt.alt || []
+  this.__cache__ = null
+}
+
+/**
+ * Ruler.before(beforeName, ruleName, fn [, options])
+ * - beforeName (String): new rule will be added before this one.
+ * - ruleName (String): name of added rule.
+ * - fn (Function): rule function.
+ * - options (Object): rule options (not mandatory).
+ *
+ * Add new rule to chain before one with given name. See also
+ * [[Ruler.after]], [[Ruler.push]].
+ *
+ * ##### Options:
+ *
+ * - __alt__ - array with names of "alternate" chains.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * var md = require('markdown-it')();
+ *
+ * md.block.ruler.before('paragraph', 'my_rule', function replace(state) {
+ *   //...
+ * });
+ * ```
+ **/
+Ruler.prototype.before = function (beforeName, ruleName, fn, options) {
+  const index = this.__find__(beforeName)
+  const opt = options || {}
+
+  if (index === -1) { throw new Error('Parser rule not found: ' + beforeName) }
+
+  this.__rules__.splice(index, 0, {
+    name: ruleName,
+    enabled: true,
+    fn,
+    alt: opt.alt || []
+  })
+
+  this.__cache__ = null
+}
+
+/**
+ * Ruler.after(afterName, ruleName, fn [, options])
+ * - afterName (String): new rule will be added after this one.
+ * - ruleName (String): name of added rule.
+ * - fn (Function): rule function.
+ * - options (Object): rule options (not mandatory).
+ *
+ * Add new rule to chain after one with given name. See also
+ * [[Ruler.before]], [[Ruler.push]].
+ *
+ * ##### Options:
+ *
+ * - __alt__ - array with names of "alternate" chains.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * var md = require('markdown-it')();
+ *
+ * md.inline.ruler.after('text', 'my_rule', function replace(state) {
+ *   //...
+ * });
+ * ```
+ **/
+Ruler.prototype.after = function (afterName, ruleName, fn, options) {
+  const index = this.__find__(afterName)
+  const opt = options || {}
+
+  if (index === -1) { throw new Error('Parser rule not found: ' + afterName) }
+
+  this.__rules__.splice(index + 1, 0, {
+    name: ruleName,
+    enabled: true,
+    fn,
+    alt: opt.alt || []
+  })
+
+  this.__cache__ = null
+}
+
+/**
+ * Ruler.push(ruleName, fn [, options])
+ * - ruleName (String): name of added rule.
+ * - fn (Function): rule function.
+ * - options (Object): rule options (not mandatory).
+ *
+ * Push new rule to the end of chain. See also
+ * [[Ruler.before]], [[Ruler.after]].
+ *
+ * ##### Options:
+ *
+ * - __alt__ - array with names of "alternate" chains.
+ *
+ * ##### Example
+ *
+ * ```javascript
+ * var md = require('markdown-it')();
+ *
+ * md.core.ruler.push('my_rule', function replace(state) {
+ *   //...
+ * });
+ * ```
+ **/
+Ruler.prototype.push = function (ruleName, fn, options) {
+  const opt = options || {}
+
+  this.__rules__.push({
+    name: ruleName,
+    enabled: true,
+    fn,
+    alt: opt.alt || []
+  })
+
+  this.__cache__ = null
+}
+
+/**
+ * Ruler.enable(list [, ignoreInvalid]) -> Array
+ * - list (String|Array): list of rule names to enable.
+ * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found.
+ *
+ * Enable rules with given names. If any rule name not found - throw Error.
+ * Errors can be disabled by second param.
+ *
+ * Returns list of found rule names (if no exception happened).
+ *
+ * See also [[Ruler.disable]], [[Ruler.enableOnly]].
+ **/
+Ruler.prototype.enable = function (list, ignoreInvalid) {
+  if (!Array.isArray(list)) { list = [list] }
+
+  const result = []
+
+  // Search by name and enable
+  list.forEach(function (name) {
+    const idx = this.__find__(name)
+
+    if (idx < 0) {
+      if (ignoreInvalid) { return }
+      throw new Error('Rules manager: invalid rule name ' + name)
+    }
+    this.__rules__[idx].enabled = true
+    result.push(name)
+  }, this)
+
+  this.__cache__ = null
+  return result
+}
+
+/**
+ * Ruler.enableOnly(list [, ignoreInvalid])
+ * - list (String|Array): list of rule names to enable (whitelist).
+ * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found.
+ *
+ * Enable rules with given names, and disable everything else. If any rule name
+ * not found - throw Error. Errors can be disabled by second param.
+ *
+ * See also [[Ruler.disable]], [[Ruler.enable]].
+ **/
+Ruler.prototype.enableOnly = function (list, ignoreInvalid) {
+  if (!Array.isArray(list)) { list = [list] }
+
+  this.__rules__.forEach(function (rule) { rule.enabled = false })
+
+  this.enable(list, ignoreInvalid)
+}
+
+/**
+ * Ruler.disable(list [, ignoreInvalid]) -> Array
+ * - list (String|Array): list of rule names to disable.
+ * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found.
+ *
+ * Disable rules with given names. If any rule name not found - throw Error.
+ * Errors can be disabled by second param.
+ *
+ * Returns list of found rule names (if no exception happened).
+ *
+ * See also [[Ruler.enable]], [[Ruler.enableOnly]].
+ **/
+Ruler.prototype.disable = function (list, ignoreInvalid) {
+  if (!Array.isArray(list)) { list = [list] }
+
+  const result = []
+
+  // Search by name and disable
+  list.forEach(function (name) {
+    const idx = this.__find__(name)
+
+    if (idx < 0) {
+      if (ignoreInvalid) { return }
+      throw new Error('Rules manager: invalid rule name ' + name)
+    }
+    this.__rules__[idx].enabled = false
+    result.push(name)
+  }, this)
+
+  this.__cache__ = null
+  return result
+}
+
+/**
+ * Ruler.getRules(chainName) -> Array
+ *
+ * Return array of active functions (rules) for given chain name. It analyzes
+ * rules configuration, compiles caches if not exists and returns result.
+ *
+ * Default chain name is `''` (empty string). It can't be skipped. That's
+ * done intentionally, to keep signature monomorphic for high speed.
+ **/
+Ruler.prototype.getRules = function (chainName) {
+  if (this.__cache__ === null) {
+    this.__compile__()
+  }
+
+  // Chain can be empty, if rules disabled. But we still have to return Array.
+  return this.__cache__[chainName] || []
+}
+
+export default Ruler

mcp-server/node_modules/markdown-it/lib/rules_block/blockquote.mjs

@@ -0,0 +1,209 @@
+// Block quotes
+
+import { isSpace } from '../common/utils.mjs'
+
+export default function blockquote (state, startLine, endLine, silent) {
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+  let max = state.eMarks[startLine]
+
+  const oldLineMax = state.lineMax
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  // check the block quote marker
+  if (state.src.charCodeAt(pos) !== 0x3E/* > */) { return false }
+
+  // we know that it's going to be a valid blockquote,
+  // so no point trying to find the end of it in silent mode
+  if (silent) { return true }
+
+  const oldBMarks  = []
+  const oldBSCount = []
+  const oldSCount  = []
+  const oldTShift  = []
+
+  const terminatorRules = state.md.block.ruler.getRules('blockquote')
+
+  const oldParentType = state.parentType
+  state.parentType = 'blockquote'
+  let lastLineEmpty = false
+  let nextLine
+
+  // Search the end of the block
+  //
+  // Block ends with either:
+  //  1. an empty line outside:
+  //     ```
+  //     > test
+  //
+  //     ```
+  //  2. an empty line inside:
+  //     ```
+  //     >
+  //     test
+  //     ```
+  //  3. another tag:
+  //     ```
+  //     > test
+  //      - - -
+  //     ```
+  for (nextLine = startLine; nextLine < endLine; nextLine++) {
+    // check if it's outdented, i.e. it's inside list item and indented
+    // less than said list item:
+    //
+    // ```
+    // 1. anything
+    //    > current blockquote
+    // 2. checking this line
+    // ```
+    const isOutdented = state.sCount[nextLine] < state.blkIndent
+
+    pos = state.bMarks[nextLine] + state.tShift[nextLine]
+    max = state.eMarks[nextLine]
+
+    if (pos >= max) {
+      // Case 1: line is not inside the blockquote, and this line is empty.
+      break
+    }
+
+    if (state.src.charCodeAt(pos++) === 0x3E/* > */ && !isOutdented) {
+      // This line is inside the blockquote.
+
+      // set offset past spaces and ">"
+      let initial = state.sCount[nextLine] + 1
+      let spaceAfterMarker
+      let adjustTab
+
+      // skip one optional space after '>'
+      if (state.src.charCodeAt(pos) === 0x20 /* space */) {
+        // ' >   test '
+        //     ^ -- position start of line here:
+        pos++
+        initial++
+        adjustTab = false
+        spaceAfterMarker = true
+      } else if (state.src.charCodeAt(pos) === 0x09 /* tab */) {
+        spaceAfterMarker = true
+
+        if ((state.bsCount[nextLine] + initial) % 4 === 3) {
+          // '  >\t  test '
+          //       ^ -- position start of line here (tab has width===1)
+          pos++
+          initial++
+          adjustTab = false
+        } else {
+          // ' >\t  test '
+          //    ^ -- position start of line here + shift bsCount slightly
+          //         to make extra space appear
+          adjustTab = true
+        }
+      } else {
+        spaceAfterMarker = false
+      }
+
+      let offset = initial
+      oldBMarks.push(state.bMarks[nextLine])
+      state.bMarks[nextLine] = pos
+
+      while (pos < max) {
+        const ch = state.src.charCodeAt(pos)
+
+        if (isSpace(ch)) {
+          if (ch === 0x09) {
+            offset += 4 - (offset + state.bsCount[nextLine] + (adjustTab ? 1 : 0)) % 4
+          } else {
+            offset++
+          }
+        } else {
+          break
+        }
+
+        pos++
+      }
+
+      lastLineEmpty = pos >= max
+
+      oldBSCount.push(state.bsCount[nextLine])
+      state.bsCount[nextLine] = state.sCount[nextLine] + 1 + (spaceAfterMarker ? 1 : 0)
+
+      oldSCount.push(state.sCount[nextLine])
+      state.sCount[nextLine] = offset - initial
+
+      oldTShift.push(state.tShift[nextLine])
+      state.tShift[nextLine] = pos - state.bMarks[nextLine]
+      continue
+    }
+
+    // Case 2: line is not inside the blockquote, and the last line was empty.
+    if (lastLineEmpty) { break }
+
+    // Case 3: another tag found.
+    let terminate = false
+    for (let i = 0, l = terminatorRules.length; i < l; i++) {
+      if (terminatorRules[i](state, nextLine, endLine, true)) {
+        terminate = true
+        break
+      }
+    }
+
+    if (terminate) {
+      // Quirk to enforce "hard termination mode" for paragraphs;
+      // normally if you call `tokenize(state, startLine, nextLine)`,
+      // paragraphs will look below nextLine for paragraph continuation,
+      // but if blockquote is terminated by another tag, they shouldn't
+      state.lineMax = nextLine
+
+      if (state.blkIndent !== 0) {
+        // state.blkIndent was non-zero, we now set it to zero,
+        // so we need to re-calculate all offsets to appear as
+        // if indent wasn't changed
+        oldBMarks.push(state.bMarks[nextLine])
+        oldBSCount.push(state.bsCount[nextLine])
+        oldTShift.push(state.tShift[nextLine])
+        oldSCount.push(state.sCount[nextLine])
+        state.sCount[nextLine] -= state.blkIndent
+      }
+
+      break
+    }
+
+    oldBMarks.push(state.bMarks[nextLine])
+    oldBSCount.push(state.bsCount[nextLine])
+    oldTShift.push(state.tShift[nextLine])
+    oldSCount.push(state.sCount[nextLine])
+
+    // A negative indentation means that this is a paragraph continuation
+    //
+    state.sCount[nextLine] = -1
+  }
+
+  const oldIndent = state.blkIndent
+  state.blkIndent = 0
+
+  const token_o  = state.push('blockquote_open', 'blockquote', 1)
+  token_o.markup = '>'
+  const lines = [startLine, 0]
+  token_o.map    = lines
+
+  state.md.block.tokenize(state, startLine, nextLine)
+
+  const token_c  = state.push('blockquote_close', 'blockquote', -1)
+  token_c.markup = '>'
+
+  state.lineMax = oldLineMax
+  state.parentType = oldParentType
+  lines[1] = state.line
+
+  // Restore original tShift; this might not be necessary since the parser
+  // has already been here, but just to make sure we can do that.
+  for (let i = 0; i < oldTShift.length; i++) {
+    state.bMarks[i + startLine] = oldBMarks[i]
+    state.tShift[i + startLine] = oldTShift[i]
+    state.sCount[i + startLine] = oldSCount[i]
+    state.bsCount[i + startLine] = oldBSCount[i]
+  }
+  state.blkIndent = oldIndent
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/code.mjs

@@ -0,0 +1,30 @@
+// Code block (4 spaces padded)
+
+export default function code (state, startLine, endLine/*, silent */) {
+  if (state.sCount[startLine] - state.blkIndent < 4) { return false }
+
+  let nextLine = startLine + 1
+  let last = nextLine
+
+  while (nextLine < endLine) {
+    if (state.isEmpty(nextLine)) {
+      nextLine++
+      continue
+    }
+
+    if (state.sCount[nextLine] - state.blkIndent >= 4) {
+      nextLine++
+      last = nextLine
+      continue
+    }
+    break
+  }
+
+  state.line = last
+
+  const token   = state.push('code_block', 'code', 0)
+  token.content = state.getLines(startLine, last, 4 + state.blkIndent, false) + '\n'
+  token.map     = [startLine, state.line]
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/fence.mjs

@@ -0,0 +1,94 @@
+// fences (``` lang, ~~~ lang)
+
+export default function fence (state, startLine, endLine, silent) {
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+  let max = state.eMarks[startLine]
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  if (pos + 3 > max) { return false }
+
+  const marker = state.src.charCodeAt(pos)
+
+  if (marker !== 0x7E/* ~ */ && marker !== 0x60 /* ` */) {
+    return false
+  }
+
+  // scan marker length
+  let mem = pos
+  pos = state.skipChars(pos, marker)
+
+  let len = pos - mem
+
+  if (len < 3) { return false }
+
+  const markup = state.src.slice(mem, pos)
+  const params = state.src.slice(pos, max)
+
+  if (marker === 0x60 /* ` */) {
+    if (params.indexOf(String.fromCharCode(marker)) >= 0) {
+      return false
+    }
+  }
+
+  // Since start is found, we can report success here in validation mode
+  if (silent) { return true }
+
+  // search end of block
+  let nextLine = startLine
+  let haveEndMarker = false
+
+  for (;;) {
+    nextLine++
+    if (nextLine >= endLine) {
+      // unclosed block should be autoclosed by end of document.
+      // also block seems to be autoclosed by end of parent
+      break
+    }
+
+    pos = mem = state.bMarks[nextLine] + state.tShift[nextLine]
+    max = state.eMarks[nextLine]
+
+    if (pos < max && state.sCount[nextLine] < state.blkIndent) {
+      // non-empty line with negative indent should stop the list:
+      // - ```
+      //  test
+      break
+    }
+
+    if (state.src.charCodeAt(pos) !== marker) { continue }
+
+    if (state.sCount[nextLine] - state.blkIndent >= 4) {
+      // closing fence should be indented less than 4 spaces
+      continue
+    }
+
+    pos = state.skipChars(pos, marker)
+
+    // closing code fence must be at least as long as the opening one
+    if (pos - mem < len) { continue }
+
+    // make sure tail has spaces only
+    pos = state.skipSpaces(pos)
+
+    if (pos < max) { continue }
+
+    haveEndMarker = true
+    // found!
+    break
+  }
+
+  // If a fence has heading spaces, they should be removed from its inner block
+  len = state.sCount[startLine]
+
+  state.line = nextLine + (haveEndMarker ? 1 : 0)
+
+  const token   = state.push('fence', 'code', 0)
+  token.info    = params
+  token.content = state.getLines(startLine + 1, nextLine, len, true)
+  token.markup  = markup
+  token.map     = [startLine, state.line]
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/heading.mjs

@@ -0,0 +1,51 @@
+// heading (#, ##, ...)
+
+import { isSpace } from '../common/utils.mjs'
+
+export default function heading (state, startLine, endLine, silent) {
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+  let max = state.eMarks[startLine]
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  let ch  = state.src.charCodeAt(pos)
+
+  if (ch !== 0x23/* # */ || pos >= max) { return false }
+
+  // count heading level
+  let level = 1
+  ch = state.src.charCodeAt(++pos)
+  while (ch === 0x23/* # */ && pos < max && level <= 6) {
+    level++
+    ch = state.src.charCodeAt(++pos)
+  }
+
+  if (level > 6 || (pos < max && !isSpace(ch))) { return false }
+
+  if (silent) { return true }
+
+  // Let's cut tails like '    ###  ' from the end of string
+
+  max = state.skipSpacesBack(max, pos)
+  const tmp = state.skipCharsBack(max, 0x23, pos) // #
+  if (tmp > pos && isSpace(state.src.charCodeAt(tmp - 1))) {
+    max = tmp
+  }
+
+  state.line = startLine + 1
+
+  const token_o  = state.push('heading_open', 'h' + String(level), 1)
+  token_o.markup = '########'.slice(0, level)
+  token_o.map    = [startLine, state.line]
+
+  const token_i    = state.push('inline', '', 0)
+  token_i.content  = state.src.slice(pos, max).trim()
+  token_i.map      = [startLine, state.line]
+  token_i.children = []
+
+  const token_c  = state.push('heading_close', 'h' + String(level), -1)
+  token_c.markup = '########'.slice(0, level)
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/hr.mjs

@@ -0,0 +1,40 @@
+// Horizontal rule
+
+import { isSpace } from '../common/utils.mjs'
+
+export default function hr (state, startLine, endLine, silent) {
+  const max = state.eMarks[startLine]
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+  const marker = state.src.charCodeAt(pos++)
+
+  // Check hr marker
+  if (marker !== 0x2A/* * */ &&
+      marker !== 0x2D/* - */ &&
+      marker !== 0x5F/* _ */) {
+    return false
+  }
+
+  // markers can be mixed with spaces, but there should be at least 3 of them
+
+  let cnt = 1
+  while (pos < max) {
+    const ch = state.src.charCodeAt(pos++)
+    if (ch !== marker && !isSpace(ch)) { return false }
+    if (ch === marker) { cnt++ }
+  }
+
+  if (cnt < 3) { return false }
+
+  if (silent) { return true }
+
+  state.line = startLine + 1
+
+  const token  = state.push('hr', 'hr', 0)
+  token.map    = [startLine, state.line]
+  token.markup = Array(cnt + 1).join(String.fromCharCode(marker))
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/html_block.mjs

@@ -0,0 +1,69 @@
+// HTML block
+
+import block_names from '../common/html_blocks.mjs'
+import { HTML_OPEN_CLOSE_TAG_RE } from '../common/html_re.mjs'
+
+// An array of opening and corresponding closing sequences for html tags,
+// last argument defines whether it can terminate a paragraph or not
+//
+const HTML_SEQUENCES = [
+  [/^<(script|pre|style|textarea)(?=(\s|>|$))/i, /<\/(script|pre|style|textarea)>/i, true],
+  [/^<!--/,        /-->/,   true],
+  [/^<\?/,         /\?>/,   true],
+  [/^<![A-Z]/,     />/,     true],
+  [/^<!\[CDATA\[/, /\]\]>/, true],
+  [new RegExp('^</?(' + block_names.join('|') + ')(?=(\\s|/?>|$))', 'i'), /^$/, true],
+  [new RegExp(HTML_OPEN_CLOSE_TAG_RE.source + '\\s*$'),  /^$/, false]
+]
+
+export default function html_block (state, startLine, endLine, silent) {
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+  let max = state.eMarks[startLine]
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  if (!state.md.options.html) { return false }
+
+  if (state.src.charCodeAt(pos) !== 0x3C/* < */) { return false }
+
+  let lineText = state.src.slice(pos, max)
+
+  let i = 0
+  for (; i < HTML_SEQUENCES.length; i++) {
+    if (HTML_SEQUENCES[i][0].test(lineText)) { break }
+  }
+  if (i === HTML_SEQUENCES.length) { return false }
+
+  if (silent) {
+    // true if this sequence can be a terminator, false otherwise
+    return HTML_SEQUENCES[i][2]
+  }
+
+  let nextLine = startLine + 1
+
+  // If we are here - we detected HTML block.
+  // Let's roll down till block end.
+  if (!HTML_SEQUENCES[i][1].test(lineText)) {
+    for (; nextLine < endLine; nextLine++) {
+      if (state.sCount[nextLine] < state.blkIndent) { break }
+
+      pos = state.bMarks[nextLine] + state.tShift[nextLine]
+      max = state.eMarks[nextLine]
+      lineText = state.src.slice(pos, max)
+
+      if (HTML_SEQUENCES[i][1].test(lineText)) {
+        if (lineText.length !== 0) { nextLine++ }
+        break
+      }
+    }
+  }
+
+  state.line = nextLine
+
+  const token   = state.push('html_block', '', 0)
+  token.map     = [startLine, nextLine]
+  token.content = state.getLines(startLine, nextLine, state.blkIndent, true)
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/lheading.mjs

@@ -0,0 +1,82 @@
+// lheading (---, ===)
+
+export default function lheading (state, startLine, endLine/*, silent */) {
+  const terminatorRules = state.md.block.ruler.getRules('paragraph')
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  const oldParentType = state.parentType
+  state.parentType = 'paragraph' // use paragraph to match terminatorRules
+
+  // jump line-by-line until empty one or EOF
+  let level = 0
+  let marker
+  let nextLine = startLine + 1
+
+  for (; nextLine < endLine && !state.isEmpty(nextLine); nextLine++) {
+    // this would be a code block normally, but after paragraph
+    // it's considered a lazy continuation regardless of what's there
+    if (state.sCount[nextLine] - state.blkIndent > 3) { continue }
+
+    //
+    // Check for underline in setext header
+    //
+    if (state.sCount[nextLine] >= state.blkIndent) {
+      let pos = state.bMarks[nextLine] + state.tShift[nextLine]
+      const max = state.eMarks[nextLine]
+
+      if (pos < max) {
+        marker = state.src.charCodeAt(pos)
+
+        if (marker === 0x2D/* - */ || marker === 0x3D/* = */) {
+          pos = state.skipChars(pos, marker)
+          pos = state.skipSpaces(pos)
+
+          if (pos >= max) {
+            level = (marker === 0x3D/* = */ ? 1 : 2)
+            break
+          }
+        }
+      }
+    }
+
+    // quirk for blockquotes, this line should already be checked by that rule
+    if (state.sCount[nextLine] < 0) { continue }
+
+    // Some tags can terminate paragraph without empty line.
+    let terminate = false
+    for (let i = 0, l = terminatorRules.length; i < l; i++) {
+      if (terminatorRules[i](state, nextLine, endLine, true)) {
+        terminate = true
+        break
+      }
+    }
+    if (terminate) { break }
+  }
+
+  if (!level) {
+    // Didn't find valid underline
+    return false
+  }
+
+  const content = state.getLines(startLine, nextLine, state.blkIndent, false).trim()
+
+  state.line = nextLine + 1
+
+  const token_o    = state.push('heading_open', 'h' + String(level), 1)
+  token_o.markup   = String.fromCharCode(marker)
+  token_o.map      = [startLine, state.line]
+
+  const token_i    = state.push('inline', '', 0)
+  token_i.content  = content
+  token_i.map      = [startLine, state.line - 1]
+  token_i.children = []
+
+  const token_c    = state.push('heading_close', 'h' + String(level), -1)
+  token_c.markup   = String.fromCharCode(marker)
+
+  state.parentType = oldParentType
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/list.mjs

@@ -0,0 +1,331 @@
+// Lists
+
+import { isSpace } from '../common/utils.mjs'
+
+// Search `[-+*][\n ]`, returns next pos after marker on success
+// or -1 on fail.
+function skipBulletListMarker (state, startLine) {
+  const max = state.eMarks[startLine]
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+
+  const marker = state.src.charCodeAt(pos++)
+  // Check bullet
+  if (marker !== 0x2A/* * */ &&
+      marker !== 0x2D/* - */ &&
+      marker !== 0x2B/* + */) {
+    return -1
+  }
+
+  if (pos < max) {
+    const ch = state.src.charCodeAt(pos)
+
+    if (!isSpace(ch)) {
+      // " -test " - is not a list item
+      return -1
+    }
+  }
+
+  return pos
+}
+
+// Search `\d+[.)][\n ]`, returns next pos after marker on success
+// or -1 on fail.
+function skipOrderedListMarker (state, startLine) {
+  const start = state.bMarks[startLine] + state.tShift[startLine]
+  const max = state.eMarks[startLine]
+  let pos = start
+
+  // List marker should have at least 2 chars (digit + dot)
+  if (pos + 1 >= max) { return -1 }
+
+  let ch = state.src.charCodeAt(pos++)
+
+  if (ch < 0x30/* 0 */ || ch > 0x39/* 9 */) { return -1 }
+
+  for (;;) {
+    // EOL -> fail
+    if (pos >= max) { return -1 }
+
+    ch = state.src.charCodeAt(pos++)
+
+    if (ch >= 0x30/* 0 */ && ch <= 0x39/* 9 */) {
+      // List marker should have no more than 9 digits
+      // (prevents integer overflow in browsers)
+      if (pos - start >= 10) { return -1 }
+
+      continue
+    }
+
+    // found valid marker
+    if (ch === 0x29/* ) */ || ch === 0x2e/* . */) {
+      break
+    }
+
+    return -1
+  }
+
+  if (pos < max) {
+    ch = state.src.charCodeAt(pos)
+
+    if (!isSpace(ch)) {
+      // " 1.test " - is not a list item
+      return -1
+    }
+  }
+  return pos
+}
+
+function markTightParagraphs (state, idx) {
+  const level = state.level + 2
+
+  for (let i = idx + 2, l = state.tokens.length - 2; i < l; i++) {
+    if (state.tokens[i].level === level && state.tokens[i].type === 'paragraph_open') {
+      state.tokens[i + 2].hidden = true
+      state.tokens[i].hidden = true
+      i += 2
+    }
+  }
+}
+
+export default function list (state, startLine, endLine, silent) {
+  let max, pos, start, token
+  let nextLine = startLine
+  let tight = true
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[nextLine] - state.blkIndent >= 4) { return false }
+
+  // Special case:
+  //  - item 1
+  //   - item 2
+  //    - item 3
+  //     - item 4
+  //      - this one is a paragraph continuation
+  if (state.listIndent >= 0 &&
+      state.sCount[nextLine] - state.listIndent >= 4 &&
+      state.sCount[nextLine] < state.blkIndent) {
+    return false
+  }
+
+  let isTerminatingParagraph = false
+
+  // limit conditions when list can interrupt
+  // a paragraph (validation mode only)
+  if (silent && state.parentType === 'paragraph') {
+    // Next list item should still terminate previous list item;
+    //
+    // This code can fail if plugins use blkIndent as well as lists,
+    // but I hope the spec gets fixed long before that happens.
+    //
+    if (state.sCount[nextLine] >= state.blkIndent) {
+      isTerminatingParagraph = true
+    }
+  }
+
+  // Detect list type and position after marker
+  let isOrdered
+  let markerValue
+  let posAfterMarker
+  if ((posAfterMarker = skipOrderedListMarker(state, nextLine)) >= 0) {
+    isOrdered = true
+    start = state.bMarks[nextLine] + state.tShift[nextLine]
+    markerValue = Number(state.src.slice(start, posAfterMarker - 1))
+
+    // If we're starting a new ordered list right after
+    // a paragraph, it should start with 1.
+    if (isTerminatingParagraph && markerValue !== 1) return false
+  } else if ((posAfterMarker = skipBulletListMarker(state, nextLine)) >= 0) {
+    isOrdered = false
+  } else {
+    return false
+  }
+
+  // If we're starting a new unordered list right after
+  // a paragraph, first line should not be empty.
+  if (isTerminatingParagraph) {
+    if (state.skipSpaces(posAfterMarker) >= state.eMarks[nextLine]) return false
+  }
+
+  // For validation mode we can terminate immediately
+  if (silent) { return true }
+
+  // We should terminate list on style change. Remember first one to compare.
+  const markerCharCode = state.src.charCodeAt(posAfterMarker - 1)
+
+  // Start list
+  const listTokIdx = state.tokens.length
+
+  if (isOrdered) {
+    token       = state.push('ordered_list_open', 'ol', 1)
+    if (markerValue !== 1) {
+      token.attrs = [['start', markerValue]]
+    }
+  } else {
+    token       = state.push('bullet_list_open', 'ul', 1)
+  }
+
+  const listLines = [nextLine, 0]
+  token.map    = listLines
+  token.markup = String.fromCharCode(markerCharCode)
+
+  //
+  // Iterate list items
+  //
+
+  let prevEmptyEnd = false
+  const terminatorRules = state.md.block.ruler.getRules('list')
+
+  const oldParentType = state.parentType
+  state.parentType = 'list'
+
+  while (nextLine < endLine) {
+    pos = posAfterMarker
+    max = state.eMarks[nextLine]
+
+    const initial = state.sCount[nextLine] + posAfterMarker - (state.bMarks[nextLine] + state.tShift[nextLine])
+    let offset = initial
+
+    while (pos < max) {
+      const ch = state.src.charCodeAt(pos)
+
+      if (ch === 0x09) {
+        offset += 4 - (offset + state.bsCount[nextLine]) % 4
+      } else if (ch === 0x20) {
+        offset++
+      } else {
+        break
+      }
+
+      pos++
+    }
+
+    const contentStart = pos
+    let indentAfterMarker
+
+    if (contentStart >= max) {
+      // trimming space in "-    \n  3" case, indent is 1 here
+      indentAfterMarker = 1
+    } else {
+      indentAfterMarker = offset - initial
+    }
+
+    // If we have more than 4 spaces, the indent is 1
+    // (the rest is just indented code block)
+    if (indentAfterMarker > 4) { indentAfterMarker = 1 }
+
+    // "  -  test"
+    //  ^^^^^ - calculating total length of this thing
+    const indent = initial + indentAfterMarker
+
+    // Run subparser & write tokens
+    token        = state.push('list_item_open', 'li', 1)
+    token.markup = String.fromCharCode(markerCharCode)
+    const itemLines = [nextLine, 0]
+    token.map    = itemLines
+    if (isOrdered) {
+      token.info = state.src.slice(start, posAfterMarker - 1)
+    }
+
+    // change current state, then restore it after parser subcall
+    const oldTight = state.tight
+    const oldTShift = state.tShift[nextLine]
+    const oldSCount = state.sCount[nextLine]
+
+    //  - example list
+    // ^ listIndent position will be here
+    //   ^ blkIndent position will be here
+    //
+    const oldListIndent = state.listIndent
+    state.listIndent = state.blkIndent
+    state.blkIndent = indent
+
+    state.tight = true
+    state.tShift[nextLine] = contentStart - state.bMarks[nextLine]
+    state.sCount[nextLine] = offset
+
+    if (contentStart >= max && state.isEmpty(nextLine + 1)) {
+      // workaround for this case
+      // (list item is empty, list terminates before "foo"):
+      // ~~~~~~~~
+      //   -
+      //
+      //     foo
+      // ~~~~~~~~
+      state.line = Math.min(state.line + 2, endLine)
+    } else {
+      state.md.block.tokenize(state, nextLine, endLine, true)
+    }
+
+    // If any of list item is tight, mark list as tight
+    if (!state.tight || prevEmptyEnd) {
+      tight = false
+    }
+    // Item become loose if finish with empty line,
+    // but we should filter last element, because it means list finish
+    prevEmptyEnd = (state.line - nextLine) > 1 && state.isEmpty(state.line - 1)
+
+    state.blkIndent = state.listIndent
+    state.listIndent = oldListIndent
+    state.tShift[nextLine] = oldTShift
+    state.sCount[nextLine] = oldSCount
+    state.tight = oldTight
+
+    token        = state.push('list_item_close', 'li', -1)
+    token.markup = String.fromCharCode(markerCharCode)
+
+    nextLine = state.line
+    itemLines[1] = nextLine
+
+    if (nextLine >= endLine) { break }
+
+    //
+    // Try to check if list is terminated or continued.
+    //
+    if (state.sCount[nextLine] < state.blkIndent) { break }
+
+    // if it's indented more than 3 spaces, it should be a code block
+    if (state.sCount[nextLine] - state.blkIndent >= 4) { break }
+
+    // fail if terminating block found
+    let terminate = false
+    for (let i = 0, l = terminatorRules.length; i < l; i++) {
+      if (terminatorRules[i](state, nextLine, endLine, true)) {
+        terminate = true
+        break
+      }
+    }
+    if (terminate) { break }
+
+    // fail if list has another type
+    if (isOrdered) {
+      posAfterMarker = skipOrderedListMarker(state, nextLine)
+      if (posAfterMarker < 0) { break }
+      start = state.bMarks[nextLine] + state.tShift[nextLine]
+    } else {
+      posAfterMarker = skipBulletListMarker(state, nextLine)
+      if (posAfterMarker < 0) { break }
+    }
+
+    if (markerCharCode !== state.src.charCodeAt(posAfterMarker - 1)) { break }
+  }
+
+  // Finalize list
+  if (isOrdered) {
+    token = state.push('ordered_list_close', 'ol', -1)
+  } else {
+    token = state.push('bullet_list_close', 'ul', -1)
+  }
+  token.markup = String.fromCharCode(markerCharCode)
+
+  listLines[1] = nextLine
+  state.line = nextLine
+
+  state.parentType = oldParentType
+
+  // mark paragraphs tight if needed
+  if (tight) {
+    markTightParagraphs(state, listTokIdx)
+  }
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/paragraph.mjs

@@ -0,0 +1,46 @@
+// Paragraph
+
+export default function paragraph (state, startLine, endLine) {
+  const terminatorRules = state.md.block.ruler.getRules('paragraph')
+  const oldParentType = state.parentType
+  let nextLine = startLine + 1
+  state.parentType = 'paragraph'
+
+  // jump line-by-line until empty one or EOF
+  for (; nextLine < endLine && !state.isEmpty(nextLine); nextLine++) {
+    // this would be a code block normally, but after paragraph
+    // it's considered a lazy continuation regardless of what's there
+    if (state.sCount[nextLine] - state.blkIndent > 3) { continue }
+
+    // quirk for blockquotes, this line should already be checked by that rule
+    if (state.sCount[nextLine] < 0) { continue }
+
+    // Some tags can terminate paragraph without empty line.
+    let terminate = false
+    for (let i = 0, l = terminatorRules.length; i < l; i++) {
+      if (terminatorRules[i](state, nextLine, endLine, true)) {
+        terminate = true
+        break
+      }
+    }
+    if (terminate) { break }
+  }
+
+  const content = state.getLines(startLine, nextLine, state.blkIndent, false).trim()
+
+  state.line = nextLine
+
+  const token_o    = state.push('paragraph_open', 'p', 1)
+  token_o.map      = [startLine, state.line]
+
+  const token_i    = state.push('inline', '', 0)
+  token_i.content  = content
+  token_i.map      = [startLine, state.line]
+  token_i.children = []
+
+  state.push('paragraph_close', 'p', -1)
+
+  state.parentType = oldParentType
+
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/reference.mjs

@@ -0,0 +1,212 @@
+import { isSpace, normalizeReference } from '../common/utils.mjs'
+
+export default function reference (state, startLine, _endLine, silent) {
+  let pos = state.bMarks[startLine] + state.tShift[startLine]
+  let max = state.eMarks[startLine]
+  let nextLine = startLine + 1
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+
+  if (state.src.charCodeAt(pos) !== 0x5B/* [ */) { return false }
+
+  function getNextLine (nextLine) {
+    const endLine = state.lineMax
+
+    if (nextLine >= endLine || state.isEmpty(nextLine)) {
+      // empty line or end of input
+      return null
+    }
+
+    let isContinuation = false
+
+    // this would be a code block normally, but after paragraph
+    // it's considered a lazy continuation regardless of what's there
+    if (state.sCount[nextLine] - state.blkIndent > 3) { isContinuation = true }
+
+    // quirk for blockquotes, this line should already be checked by that rule
+    if (state.sCount[nextLine] < 0) { isContinuation = true }
+
+    if (!isContinuation) {
+      const terminatorRules = state.md.block.ruler.getRules('reference')
+      const oldParentType = state.parentType
+      state.parentType = 'reference'
+
+      // Some tags can terminate paragraph without empty line.
+      let terminate = false
+      for (let i = 0, l = terminatorRules.length; i < l; i++) {
+        if (terminatorRules[i](state, nextLine, endLine, true)) {
+          terminate = true
+          break
+        }
+      }
+
+      state.parentType = oldParentType
+      if (terminate) {
+        // terminated by another block
+        return null
+      }
+    }
+
+    const pos = state.bMarks[nextLine] + state.tShift[nextLine]
+    const max = state.eMarks[nextLine]
+
+    // max + 1 explicitly includes the newline
+    return state.src.slice(pos, max + 1)
+  }
+
+  let str = state.src.slice(pos, max + 1)
+
+  max = str.length
+  let labelEnd = -1
+
+  for (pos = 1; pos < max; pos++) {
+    const ch = str.charCodeAt(pos)
+    if (ch === 0x5B /* [ */) {
+      return false
+    } else if (ch === 0x5D /* ] */) {
+      labelEnd = pos
+      break
+    } else if (ch === 0x0A /* \n */) {
+      const lineContent = getNextLine(nextLine)
+      if (lineContent !== null) {
+        str += lineContent
+        max = str.length
+        nextLine++
+      }
+    } else if (ch === 0x5C /* \ */) {
+      pos++
+      if (pos < max && str.charCodeAt(pos) === 0x0A) {
+        const lineContent = getNextLine(nextLine)
+        if (lineContent !== null) {
+          str += lineContent
+          max = str.length
+          nextLine++
+        }
+      }
+    }
+  }
+
+  if (labelEnd < 0 || str.charCodeAt(labelEnd + 1) !== 0x3A/* : */) { return false }
+
+  // [label]:   destination   'title'
+  //         ^^^ skip optional whitespace here
+  for (pos = labelEnd + 2; pos < max; pos++) {
+    const ch = str.charCodeAt(pos)
+    if (ch === 0x0A) {
+      const lineContent = getNextLine(nextLine)
+      if (lineContent !== null) {
+        str += lineContent
+        max = str.length
+        nextLine++
+      }
+    } else if (isSpace(ch)) {
+      /* eslint no-empty:0 */
+    } else {
+      break
+    }
+  }
+
+  // [label]:   destination   'title'
+  //            ^^^^^^^^^^^ parse this
+  const destRes = state.md.helpers.parseLinkDestination(str, pos, max)
+  if (!destRes.ok) { return false }
+
+  const href = state.md.normalizeLink(destRes.str)
+  if (!state.md.validateLink(href)) { return false }
+
+  pos = destRes.pos
+
+  // save cursor state, we could require to rollback later
+  const destEndPos = pos
+  const destEndLineNo = nextLine
+
+  // [label]:   destination   'title'
+  //                       ^^^ skipping those spaces
+  const start = pos
+  for (; pos < max; pos++) {
+    const ch = str.charCodeAt(pos)
+    if (ch === 0x0A) {
+      const lineContent = getNextLine(nextLine)
+      if (lineContent !== null) {
+        str += lineContent
+        max = str.length
+        nextLine++
+      }
+    } else if (isSpace(ch)) {
+      /* eslint no-empty:0 */
+    } else {
+      break
+    }
+  }
+
+  // [label]:   destination   'title'
+  //                          ^^^^^^^ parse this
+  let titleRes = state.md.helpers.parseLinkTitle(str, pos, max)
+  while (titleRes.can_continue) {
+    const lineContent = getNextLine(nextLine)
+    if (lineContent === null) break
+    str += lineContent
+    pos = max
+    max = str.length
+    nextLine++
+    titleRes = state.md.helpers.parseLinkTitle(str, pos, max, titleRes)
+  }
+  let title
+
+  if (pos < max && start !== pos && titleRes.ok) {
+    title = titleRes.str
+    pos = titleRes.pos
+  } else {
+    title = ''
+    pos = destEndPos
+    nextLine = destEndLineNo
+  }
+
+  // skip trailing spaces until the rest of the line
+  while (pos < max) {
+    const ch = str.charCodeAt(pos)
+    if (!isSpace(ch)) { break }
+    pos++
+  }
+
+  if (pos < max && str.charCodeAt(pos) !== 0x0A) {
+    if (title) {
+      // garbage at the end of the line after title,
+      // but it could still be a valid reference if we roll back
+      title = ''
+      pos = destEndPos
+      nextLine = destEndLineNo
+      while (pos < max) {
+        const ch = str.charCodeAt(pos)
+        if (!isSpace(ch)) { break }
+        pos++
+      }
+    }
+  }
+
+  if (pos < max && str.charCodeAt(pos) !== 0x0A) {
+    // garbage at the end of the line
+    return false
+  }
+
+  const label = normalizeReference(str.slice(1, labelEnd))
+  if (!label) {
+    // CommonMark 0.20 disallows empty labels
+    return false
+  }
+
+  // Reference can not terminate anything. This check is for safety only.
+  /* istanbul ignore if */
+  if (silent) { return true }
+
+  if (typeof state.env.references === 'undefined') {
+    state.env.references = {}
+  }
+  if (typeof state.env.references[label] === 'undefined') {
+    state.env.references[label] = { title, href }
+  }
+
+  state.line = nextLine
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_block/state_block.mjs

@@ -0,0 +1,220 @@
+// Parser state class
+
+import Token from '../token.mjs'
+import { isSpace } from '../common/utils.mjs'
+
+function StateBlock (src, md, env, tokens) {
+  this.src = src
+
+  // link to parser instance
+  this.md     = md
+
+  this.env = env
+
+  //
+  // Internal state vartiables
+  //
+
+  this.tokens = tokens
+
+  this.bMarks = []  // line begin offsets for fast jumps
+  this.eMarks = []  // line end offsets for fast jumps
+  this.tShift = []  // offsets of the first non-space characters (tabs not expanded)
+  this.sCount = []  // indents for each line (tabs expanded)
+
+  // An amount of virtual spaces (tabs expanded) between beginning
+  // of each line (bMarks) and real beginning of that line.
+  //
+  // It exists only as a hack because blockquotes override bMarks
+  // losing information in the process.
+  //
+  // It's used only when expanding tabs, you can think about it as
+  // an initial tab length, e.g. bsCount=21 applied to string `\t123`
+  // means first tab should be expanded to 4-21%4 === 3 spaces.
+  //
+  this.bsCount = []
+
+  // block parser variables
+
+  // required block content indent (for example, if we are
+  // inside a list, it would be positioned after list marker)
+  this.blkIndent  = 0
+  this.line       = 0 // line index in src
+  this.lineMax    = 0 // lines count
+  this.tight      = false  // loose/tight mode for lists
+  this.ddIndent   = -1 // indent of the current dd block (-1 if there isn't any)
+  this.listIndent = -1 // indent of the current list block (-1 if there isn't any)
+
+  // can be 'blockquote', 'list', 'root', 'paragraph' or 'reference'
+  // used in lists to determine if they interrupt a paragraph
+  this.parentType = 'root'
+
+  this.level = 0
+
+  // Create caches
+  // Generate markers.
+  const s = this.src
+
+  for (let start = 0, pos = 0, indent = 0, offset = 0, len = s.length, indent_found = false; pos < len; pos++) {
+    const ch = s.charCodeAt(pos)
+
+    if (!indent_found) {
+      if (isSpace(ch)) {
+        indent++
+
+        if (ch === 0x09) {
+          offset += 4 - offset % 4
+        } else {
+          offset++
+        }
+        continue
+      } else {
+        indent_found = true
+      }
+    }
+
+    if (ch === 0x0A || pos === len - 1) {
+      if (ch !== 0x0A) { pos++ }
+      this.bMarks.push(start)
+      this.eMarks.push(pos)
+      this.tShift.push(indent)
+      this.sCount.push(offset)
+      this.bsCount.push(0)
+
+      indent_found = false
+      indent = 0
+      offset = 0
+      start = pos + 1
+    }
+  }
+
+  // Push fake entry to simplify cache bounds checks
+  this.bMarks.push(s.length)
+  this.eMarks.push(s.length)
+  this.tShift.push(0)
+  this.sCount.push(0)
+  this.bsCount.push(0)
+
+  this.lineMax = this.bMarks.length - 1 // don't count last fake line
+}
+
+// Push new token to "stream".
+//
+StateBlock.prototype.push = function (type, tag, nesting) {
+  const token = new Token(type, tag, nesting)
+  token.block = true
+
+  if (nesting < 0) this.level-- // closing tag
+  token.level = this.level
+  if (nesting > 0) this.level++ // opening tag
+
+  this.tokens.push(token)
+  return token
+}
+
+StateBlock.prototype.isEmpty = function isEmpty (line) {
+  return this.bMarks[line] + this.tShift[line] >= this.eMarks[line]
+}
+
+StateBlock.prototype.skipEmptyLines = function skipEmptyLines (from) {
+  for (let max = this.lineMax; from < max; from++) {
+    if (this.bMarks[from] + this.tShift[from] < this.eMarks[from]) {
+      break
+    }
+  }
+  return from
+}
+
+// Skip spaces from given position.
+StateBlock.prototype.skipSpaces = function skipSpaces (pos) {
+  for (let max = this.src.length; pos < max; pos++) {
+    const ch = this.src.charCodeAt(pos)
+    if (!isSpace(ch)) { break }
+  }
+  return pos
+}
+
+// Skip spaces from given position in reverse.
+StateBlock.prototype.skipSpacesBack = function skipSpacesBack (pos, min) {
+  if (pos <= min) { return pos }
+
+  while (pos > min) {
+    if (!isSpace(this.src.charCodeAt(--pos))) { return pos + 1 }
+  }
+  return pos
+}
+
+// Skip char codes from given position
+StateBlock.prototype.skipChars = function skipChars (pos, code) {
+  for (let max = this.src.length; pos < max; pos++) {
+    if (this.src.charCodeAt(pos) !== code) { break }
+  }
+  return pos
+}
+
+// Skip char codes reverse from given position - 1
+StateBlock.prototype.skipCharsBack = function skipCharsBack (pos, code, min) {
+  if (pos <= min) { return pos }
+
+  while (pos > min) {
+    if (code !== this.src.charCodeAt(--pos)) { return pos + 1 }
+  }
+  return pos
+}
+
+// cut lines range from source.
+StateBlock.prototype.getLines = function getLines (begin, end, indent, keepLastLF) {
+  if (begin >= end) {
+    return ''
+  }
+
+  const queue = new Array(end - begin)
+
+  for (let i = 0, line = begin; line < end; line++, i++) {
+    let lineIndent = 0
+    const lineStart = this.bMarks[line]
+    let first = lineStart
+    let last
+
+    if (line + 1 < end || keepLastLF) {
+      // No need for bounds check because we have fake entry on tail.
+      last = this.eMarks[line] + 1
+    } else {
+      last = this.eMarks[line]
+    }
+
+    while (first < last && lineIndent < indent) {
+      const ch = this.src.charCodeAt(first)
+
+      if (isSpace(ch)) {
+        if (ch === 0x09) {
+          lineIndent += 4 - (lineIndent + this.bsCount[line]) % 4
+        } else {
+          lineIndent++
+        }
+      } else if (first - lineStart < this.tShift[line]) {
+        // patched tShift masked characters to look like spaces (blockquotes, list markers)
+        lineIndent++
+      } else {
+        break
+      }
+
+      first++
+    }
+
+    if (lineIndent > indent) {
+      // partially expanding tabs in code blocks, e.g '\t\tfoobar'
+      // with indent=2 becomes '  \tfoobar'
+      queue[i] = new Array(lineIndent - indent + 1).join(' ') + this.src.slice(first, last)
+    } else {
+      queue[i] = this.src.slice(first, last)
+    }
+  }
+
+  return queue.join('')
+}
+
+// re-export Token class to use in block rules
+StateBlock.prototype.Token = Token
+
+export default StateBlock

mcp-server/node_modules/markdown-it/lib/rules_block/table.mjs

@@ -0,0 +1,228 @@
+// GFM table, https://github.github.com/gfm/#tables-extension-
+
+import { isSpace } from '../common/utils.mjs'
+
+// Limit the amount of empty autocompleted cells in a table,
+// see https://github.com/markdown-it/markdown-it/issues/1000,
+//
+// Both pulldown-cmark and commonmark-hs limit the number of cells this way to ~200k.
+// We set it to 65k, which can expand user input by a factor of x370
+// (256x256 square is 1.8kB expanded into 650kB).
+const MAX_AUTOCOMPLETED_CELLS = 0x10000
+
+function getLine (state, line) {
+  const pos = state.bMarks[line] + state.tShift[line]
+  const max = state.eMarks[line]
+
+  return state.src.slice(pos, max)
+}
+
+function escapedSplit (str) {
+  const result = []
+  const max = str.length
+
+  let pos = 0
+  let ch = str.charCodeAt(pos)
+  let isEscaped = false
+  let lastPos = 0
+  let current = ''
+
+  while (pos < max) {
+    if (ch === 0x7c/* | */) {
+      if (!isEscaped) {
+        // pipe separating cells, '|'
+        result.push(current + str.substring(lastPos, pos))
+        current = ''
+        lastPos = pos + 1
+      } else {
+        // escaped pipe, '\|'
+        current += str.substring(lastPos, pos - 1)
+        lastPos = pos
+      }
+    }
+
+    isEscaped = (ch === 0x5c/* \ */)
+    pos++
+
+    ch = str.charCodeAt(pos)
+  }
+
+  result.push(current + str.substring(lastPos))
+
+  return result
+}
+
+export default function table (state, startLine, endLine, silent) {
+  // should have at least two lines
+  if (startLine + 2 > endLine) { return false }
+
+  let nextLine = startLine + 1
+
+  if (state.sCount[nextLine] < state.blkIndent) { return false }
+
+  // if it's indented more than 3 spaces, it should be a code block
+  if (state.sCount[nextLine] - state.blkIndent >= 4) { return false }
+
+  // first character of the second line should be '|', '-', ':',
+  // and no other characters are allowed but spaces;
+  // basically, this is the equivalent of /^[-:|][-:|\s]*$/ regexp
+
+  let pos = state.bMarks[nextLine] + state.tShift[nextLine]
+  if (pos >= state.eMarks[nextLine]) { return false }
+
+  const firstCh = state.src.charCodeAt(pos++)
+  if (firstCh !== 0x7C/* | */ && firstCh !== 0x2D/* - */ && firstCh !== 0x3A/* : */) { return false }
+
+  if (pos >= state.eMarks[nextLine]) { return false }
+
+  const secondCh = state.src.charCodeAt(pos++)
+  if (secondCh !== 0x7C/* | */ && secondCh !== 0x2D/* - */ && secondCh !== 0x3A/* : */ && !isSpace(secondCh)) {
+    return false
+  }
+
+  // if first character is '-', then second character must not be a space
+  // (due to parsing ambiguity with list)
+  if (firstCh === 0x2D/* - */ && isSpace(secondCh)) { return false }
+
+  while (pos < state.eMarks[nextLine]) {
+    const ch = state.src.charCodeAt(pos)
+
+    if (ch !== 0x7C/* | */ && ch !== 0x2D/* - */ && ch !== 0x3A/* : */ && !isSpace(ch)) { return false }
+
+    pos++
+  }
+
+  let lineText = getLine(state, startLine + 1)
+  let columns = lineText.split('|')
+  const aligns = []
+  for (let i = 0; i < columns.length; i++) {
+    const t = columns[i].trim()
+    if (!t) {
+      // allow empty columns before and after table, but not in between columns;
+      // e.g. allow ` |---| `, disallow ` ---||--- `
+      if (i === 0 || i === columns.length - 1) {
+        continue
+      } else {
+        return false
+      }
+    }
+
+    if (!/^:?-+:?$/.test(t)) { return false }
+    if (t.charCodeAt(t.length - 1) === 0x3A/* : */) {
+      aligns.push(t.charCodeAt(0) === 0x3A/* : */ ? 'center' : 'right')
+    } else if (t.charCodeAt(0) === 0x3A/* : */) {
+      aligns.push('left')
+    } else {
+      aligns.push('')
+    }
+  }
+
+  lineText = getLine(state, startLine).trim()
+  if (lineText.indexOf('|') === -1) { return false }
+  if (state.sCount[startLine] - state.blkIndent >= 4) { return false }
+  columns = escapedSplit(lineText)
+  if (columns.length && columns[0] === '') columns.shift()
+  if (columns.length && columns[columns.length - 1] === '') columns.pop()
+
+  // header row will define an amount of columns in the entire table,
+  // and align row should be exactly the same (the rest of the rows can differ)
+  const columnCount = columns.length
+  if (columnCount === 0 || columnCount !== aligns.length) { return false }
+
+  if (silent) { return true }
+
+  const oldParentType = state.parentType
+  state.parentType = 'table'
+
+  // use 'blockquote' lists for termination because it's
+  // the most similar to tables
+  const terminatorRules = state.md.block.ruler.getRules('blockquote')
+
+  const token_to = state.push('table_open', 'table', 1)
+  const tableLines = [startLine, 0]
+  token_to.map = tableLines
+
+  const token_tho = state.push('thead_open', 'thead', 1)
+  token_tho.map = [startLine, startLine + 1]
+
+  const token_htro = state.push('tr_open', 'tr', 1)
+  token_htro.map = [startLine, startLine + 1]
+
+  for (let i = 0; i < columns.length; i++) {
+    const token_ho = state.push('th_open', 'th', 1)
+    if (aligns[i]) {
+      token_ho.attrs  = [['style', 'text-align:' + aligns[i]]]
+    }
+
+    const token_il = state.push('inline', '', 0)
+    token_il.content  = columns[i].trim()
+    token_il.children = []
+
+    state.push('th_close', 'th', -1)
+  }
+
+  state.push('tr_close', 'tr', -1)
+  state.push('thead_close', 'thead', -1)
+
+  let tbodyLines
+  let autocompletedCells = 0
+
+  for (nextLine = startLine + 2; nextLine < endLine; nextLine++) {
+    if (state.sCount[nextLine] < state.blkIndent) { break }
+
+    let terminate = false
+    for (let i = 0, l = terminatorRules.length; i < l; i++) {
+      if (terminatorRules[i](state, nextLine, endLine, true)) {
+        terminate = true
+        break
+      }
+    }
+
+    if (terminate) { break }
+    lineText = getLine(state, nextLine).trim()
+    if (!lineText) { break }
+    if (state.sCount[nextLine] - state.blkIndent >= 4) { break }
+    columns = escapedSplit(lineText)
+    if (columns.length && columns[0] === '') columns.shift()
+    if (columns.length && columns[columns.length - 1] === '') columns.pop()
+
+    // note: autocomplete count can be negative if user specifies more columns than header,
+    // but that does not affect intended use (which is limiting expansion)
+    autocompletedCells += columnCount - columns.length
+    if (autocompletedCells > MAX_AUTOCOMPLETED_CELLS) { break }
+
+    if (nextLine === startLine + 2) {
+      const token_tbo = state.push('tbody_open', 'tbody', 1)
+      token_tbo.map = tbodyLines = [startLine + 2, 0]
+    }
+
+    const token_tro = state.push('tr_open', 'tr', 1)
+    token_tro.map = [nextLine, nextLine + 1]
+
+    for (let i = 0; i < columnCount; i++) {
+      const token_tdo = state.push('td_open', 'td', 1)
+      if (aligns[i]) {
+        token_tdo.attrs  = [['style', 'text-align:' + aligns[i]]]
+      }
+
+      const token_il = state.push('inline', '', 0)
+      token_il.content  = columns[i] ? columns[i].trim() : ''
+      token_il.children = []
+
+      state.push('td_close', 'td', -1)
+    }
+    state.push('tr_close', 'tr', -1)
+  }
+
+  if (tbodyLines) {
+    state.push('tbody_close', 'tbody', -1)
+    tbodyLines[1] = nextLine
+  }
+
+  state.push('table_close', 'table', -1)
+  tableLines[1] = nextLine
+
+  state.parentType = oldParentType
+  state.line = nextLine
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_core/block.mjs

@@ -0,0 +1,13 @@
+export default function block (state) {
+  let token
+
+  if (state.inlineMode) {
+    token          = new state.Token('inline', '', 0)
+    token.content  = state.src
+    token.map      = [0, 1]
+    token.children = []
+    state.tokens.push(token)
+  } else {
+    state.md.block.parse(state.src, state.md, state.env, state.tokens)
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_core/inline.mjs

@@ -0,0 +1,11 @@
+export default function inline (state) {
+  const tokens = state.tokens
+
+  // Parse inlines
+  for (let i = 0, l = tokens.length; i < l; i++) {
+    const tok = tokens[i]
+    if (tok.type === 'inline') {
+      state.md.inline.parse(tok.content, state.md, state.env, tok.children)
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_core/linkify.mjs

@@ -0,0 +1,134 @@
+// Replace link-like texts with link nodes.
+//
+// Currently restricted by `md.validateLink()` to http/https/ftp
+//
+
+import { arrayReplaceAt } from '../common/utils.mjs'
+
+function isLinkOpen (str) {
+  return /^<a[>\s]/i.test(str)
+}
+function isLinkClose (str) {
+  return /^<\/a\s*>/i.test(str)
+}
+
+export default function linkify (state) {
+  const blockTokens = state.tokens
+
+  if (!state.md.options.linkify) { return }
+
+  for (let j = 0, l = blockTokens.length; j < l; j++) {
+    if (blockTokens[j].type !== 'inline' ||
+        !state.md.linkify.pretest(blockTokens[j].content)) {
+      continue
+    }
+
+    let tokens = blockTokens[j].children
+
+    let htmlLinkLevel = 0
+
+    // We scan from the end, to keep position when new tags added.
+    // Use reversed logic in links start/end match
+    for (let i = tokens.length - 1; i >= 0; i--) {
+      const currentToken = tokens[i]
+
+      // Skip content of markdown links
+      if (currentToken.type === 'link_close') {
+        i--
+        while (tokens[i].level !== currentToken.level && tokens[i].type !== 'link_open') {
+          i--
+        }
+        continue
+      }
+
+      // Skip content of html tag links
+      if (currentToken.type === 'html_inline') {
+        if (isLinkOpen(currentToken.content) && htmlLinkLevel > 0) {
+          htmlLinkLevel--
+        }
+        if (isLinkClose(currentToken.content)) {
+          htmlLinkLevel++
+        }
+      }
+      if (htmlLinkLevel > 0) { continue }
+
+      if (currentToken.type === 'text' && state.md.linkify.test(currentToken.content)) {
+        const text = currentToken.content
+        let links = state.md.linkify.match(text)
+
+        // Now split string to nodes
+        const nodes = []
+        let level = currentToken.level
+        let lastPos = 0
+
+        // forbid escape sequence at the start of the string,
+        // this avoids http\://example.com/ from being linkified as
+        // http:<a href="//example.com/">//example.com/</a>
+        if (links.length > 0 &&
+            links[0].index === 0 &&
+            i > 0 &&
+            tokens[i - 1].type === 'text_special') {
+          links = links.slice(1)
+        }
+
+        for (let ln = 0; ln < links.length; ln++) {
+          const url = links[ln].url
+          const fullUrl = state.md.normalizeLink(url)
+          if (!state.md.validateLink(fullUrl)) { continue }
+
+          let urlText = links[ln].text
+
+          // Linkifier might send raw hostnames like "example.com", where url
+          // starts with domain name. So we prepend http:// in those cases,
+          // and remove it afterwards.
+          //
+          if (!links[ln].schema) {
+            urlText = state.md.normalizeLinkText('http://' + urlText).replace(/^http:\/\//, '')
+          } else if (links[ln].schema === 'mailto:' && !/^mailto:/i.test(urlText)) {
+            urlText = state.md.normalizeLinkText('mailto:' + urlText).replace(/^mailto:/, '')
+          } else {
+            urlText = state.md.normalizeLinkText(urlText)
+          }
+
+          const pos = links[ln].index
+
+          if (pos > lastPos) {
+            const token   = new state.Token('text', '', 0)
+            token.content = text.slice(lastPos, pos)
+            token.level   = level
+            nodes.push(token)
+          }
+
+          const token_o   = new state.Token('link_open', 'a', 1)
+          token_o.attrs   = [['href', fullUrl]]
+          token_o.level   = level++
+          token_o.markup  = 'linkify'
+          token_o.info    = 'auto'
+          nodes.push(token_o)
+
+          const token_t   = new state.Token('text', '', 0)
+          token_t.content = urlText
+          token_t.level   = level
+          nodes.push(token_t)
+
+          const token_c   = new state.Token('link_close', 'a', -1)
+          token_c.level   = --level
+          token_c.markup  = 'linkify'
+          token_c.info    = 'auto'
+          nodes.push(token_c)
+
+          lastPos = links[ln].lastIndex
+        }
+        if (lastPos < text.length) {
+          const token   = new state.Token('text', '', 0)
+          token.content = text.slice(lastPos)
+          token.level   = level
+          nodes.push(token)
+        }
+
+        // replace current node
+        blockTokens[j].children = tokens = arrayReplaceAt(tokens, i, nodes)
+      }
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_core/normalize.mjs

@@ -0,0 +1,17 @@
+// Normalize input string
+
+// https://spec.commonmark.org/0.29/#line-ending
+const NEWLINES_RE  = /\r\n?|\n/g
+const NULL_RE      = /\0/g
+
+export default function normalize (state) {
+  let str
+
+  // Normalize newlines
+  str = state.src.replace(NEWLINES_RE, '\n')
+
+  // Replace NULL characters
+  str = str.replace(NULL_RE, '\uFFFD')
+
+  state.src = str
+}

mcp-server/node_modules/markdown-it/lib/rules_core/replacements.mjs

@@ -0,0 +1,101 @@
+// Simple typographic replacements
+//
+// (c) (C) → ©
+// (tm) (TM) → ™
+// (r) (R) → ®
+// +- → ±
+// ... → … (also ?.... → ?.., !.... → !..)
+// ???????? → ???, !!!!! → !!!, `,,` → `,`
+// -- → –, --- → —
+//
+
+// TODO:
+// - fractionals 1/2, 1/4, 3/4 -> ½, ¼, ¾
+// - multiplications 2 x 4 -> 2 × 4
+
+const RARE_RE = /\+-|\.\.|\?\?\?\?|!!!!|,,|--/
+
+// Workaround for phantomjs - need regex without /g flag,
+// or root check will fail every second time
+const SCOPED_ABBR_TEST_RE = /\((c|tm|r)\)/i
+
+const SCOPED_ABBR_RE = /\((c|tm|r)\)/ig
+const SCOPED_ABBR = {
+  c: '©',
+  r: '®',
+  tm: '™'
+}
+
+function replaceFn (match, name) {
+  return SCOPED_ABBR[name.toLowerCase()]
+}
+
+function replace_scoped (inlineTokens) {
+  let inside_autolink = 0
+
+  for (let i = inlineTokens.length - 1; i >= 0; i--) {
+    const token = inlineTokens[i]
+
+    if (token.type === 'text' && !inside_autolink) {
+      token.content = token.content.replace(SCOPED_ABBR_RE, replaceFn)
+    }
+
+    if (token.type === 'link_open' && token.info === 'auto') {
+      inside_autolink--
+    }
+
+    if (token.type === 'link_close' && token.info === 'auto') {
+      inside_autolink++
+    }
+  }
+}
+
+function replace_rare (inlineTokens) {
+  let inside_autolink = 0
+
+  for (let i = inlineTokens.length - 1; i >= 0; i--) {
+    const token = inlineTokens[i]
+
+    if (token.type === 'text' && !inside_autolink) {
+      if (RARE_RE.test(token.content)) {
+        token.content = token.content
+          .replace(/\+-/g, '±')
+          // .., ..., ....... -> …
+          // but ?..... & !..... -> ?.. & !..
+          .replace(/\.{2,}/g, '…').replace(/([?!])…/g, '$1..')
+          .replace(/([?!]){4,}/g, '$1$1$1').replace(/,{2,}/g, ',')
+          // em-dash
+          .replace(/(^|[^-])---(?=[^-]|$)/mg, '$1\u2014')
+          // en-dash
+          .replace(/(^|\s)--(?=\s|$)/mg, '$1\u2013')
+          .replace(/(^|[^-\s])--(?=[^-\s]|$)/mg, '$1\u2013')
+      }
+    }
+
+    if (token.type === 'link_open' && token.info === 'auto') {
+      inside_autolink--
+    }
+
+    if (token.type === 'link_close' && token.info === 'auto') {
+      inside_autolink++
+    }
+  }
+}
+
+export default function replace (state) {
+  let blkIdx
+
+  if (!state.md.options.typographer) { return }
+
+  for (blkIdx = state.tokens.length - 1; blkIdx >= 0; blkIdx--) {
+    if (state.tokens[blkIdx].type !== 'inline') { continue }
+
+    if (SCOPED_ABBR_TEST_RE.test(state.tokens[blkIdx].content)) {
+      replace_scoped(state.tokens[blkIdx].children)
+    }
+
+    if (RARE_RE.test(state.tokens[blkIdx].content)) {
+      replace_rare(state.tokens[blkIdx].children)
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_core/smartquotes.mjs

@@ -0,0 +1,193 @@
+// Convert straight quotation marks to typographic ones
+//
+
+import { isWhiteSpace, isPunctChar, isMdAsciiPunct } from '../common/utils.mjs'
+
+const QUOTE_TEST_RE = /['"]/
+const QUOTE_RE = /['"]/g
+const APOSTROPHE = '\u2019' /* ’ */
+
+function replaceAt (str, index, ch) {
+  return str.slice(0, index) + ch + str.slice(index + 1)
+}
+
+function process_inlines (tokens, state) {
+  let j
+
+  const stack = []
+
+  for (let i = 0; i < tokens.length; i++) {
+    const token = tokens[i]
+
+    const thisLevel = tokens[i].level
+
+    for (j = stack.length - 1; j >= 0; j--) {
+      if (stack[j].level <= thisLevel) { break }
+    }
+    stack.length = j + 1
+
+    if (token.type !== 'text') { continue }
+
+    let text = token.content
+    let pos = 0
+    let max = text.length
+
+    /* eslint no-labels:0,block-scoped-var:0 */
+    OUTER:
+    while (pos < max) {
+      QUOTE_RE.lastIndex = pos
+      const t = QUOTE_RE.exec(text)
+      if (!t) { break }
+
+      let canOpen = true
+      let canClose = true
+      pos = t.index + 1
+      const isSingle = (t[0] === "'")
+
+      // Find previous character,
+      // default to space if it's the beginning of the line
+      //
+      let lastChar = 0x20
+
+      if (t.index - 1 >= 0) {
+        lastChar = text.charCodeAt(t.index - 1)
+      } else {
+        for (j = i - 1; j >= 0; j--) {
+          if (tokens[j].type === 'softbreak' || tokens[j].type === 'hardbreak') break // lastChar defaults to 0x20
+          if (!tokens[j].content) continue // should skip all tokens except 'text', 'html_inline' or 'code_inline'
+
+          lastChar = tokens[j].content.charCodeAt(tokens[j].content.length - 1)
+          break
+        }
+      }
+
+      // Find next character,
+      // default to space if it's the end of the line
+      //
+      let nextChar = 0x20
+
+      if (pos < max) {
+        nextChar = text.charCodeAt(pos)
+      } else {
+        for (j = i + 1; j < tokens.length; j++) {
+          if (tokens[j].type === 'softbreak' || tokens[j].type === 'hardbreak') break // nextChar defaults to 0x20
+          if (!tokens[j].content) continue // should skip all tokens except 'text', 'html_inline' or 'code_inline'
+
+          nextChar = tokens[j].content.charCodeAt(0)
+          break
+        }
+      }
+
+      const isLastPunctChar = isMdAsciiPunct(lastChar) || isPunctChar(String.fromCharCode(lastChar))
+      const isNextPunctChar = isMdAsciiPunct(nextChar) || isPunctChar(String.fromCharCode(nextChar))
+
+      const isLastWhiteSpace = isWhiteSpace(lastChar)
+      const isNextWhiteSpace = isWhiteSpace(nextChar)
+
+      if (isNextWhiteSpace) {
+        canOpen = false
+      } else if (isNextPunctChar) {
+        if (!(isLastWhiteSpace || isLastPunctChar)) {
+          canOpen = false
+        }
+      }
+
+      if (isLastWhiteSpace) {
+        canClose = false
+      } else if (isLastPunctChar) {
+        if (!(isNextWhiteSpace || isNextPunctChar)) {
+          canClose = false
+        }
+      }
+
+      if (nextChar === 0x22 /* " */ && t[0] === '"') {
+        if (lastChar >= 0x30 /* 0 */ && lastChar <= 0x39 /* 9 */) {
+          // special case: 1"" - count first quote as an inch
+          canClose = canOpen = false
+        }
+      }
+
+      if (canOpen && canClose) {
+        // Replace quotes in the middle of punctuation sequence, but not
+        // in the middle of the words, i.e.:
+        //
+        // 1. foo " bar " baz - not replaced
+        // 2. foo-"-bar-"-baz - replaced
+        // 3. foo"bar"baz     - not replaced
+        //
+        canOpen = isLastPunctChar
+        canClose = isNextPunctChar
+      }
+
+      if (!canOpen && !canClose) {
+        // middle of word
+        if (isSingle) {
+          token.content = replaceAt(token.content, t.index, APOSTROPHE)
+        }
+        continue
+      }
+
+      if (canClose) {
+        // this could be a closing quote, rewind the stack to get a match
+        for (j = stack.length - 1; j >= 0; j--) {
+          let item = stack[j]
+          if (stack[j].level < thisLevel) { break }
+          if (item.single === isSingle && stack[j].level === thisLevel) {
+            item = stack[j]
+
+            let openQuote
+            let closeQuote
+            if (isSingle) {
+              openQuote = state.md.options.quotes[2]
+              closeQuote = state.md.options.quotes[3]
+            } else {
+              openQuote = state.md.options.quotes[0]
+              closeQuote = state.md.options.quotes[1]
+            }
+
+            // replace token.content *before* tokens[item.token].content,
+            // because, if they are pointing at the same token, replaceAt
+            // could mess up indices when quote length != 1
+            token.content = replaceAt(token.content, t.index, closeQuote)
+            tokens[item.token].content = replaceAt(
+              tokens[item.token].content, item.pos, openQuote)
+
+            pos += closeQuote.length - 1
+            if (item.token === i) { pos += openQuote.length - 1 }
+
+            text = token.content
+            max = text.length
+
+            stack.length = j
+            continue OUTER
+          }
+        }
+      }
+
+      if (canOpen) {
+        stack.push({
+          token: i,
+          pos: t.index,
+          single: isSingle,
+          level: thisLevel
+        })
+      } else if (canClose && isSingle) {
+        token.content = replaceAt(token.content, t.index, APOSTROPHE)
+      }
+    }
+  }
+}
+
+export default function smartquotes (state) {
+  /* eslint max-depth:0 */
+  if (!state.md.options.typographer) { return }
+
+  for (let blkIdx = state.tokens.length - 1; blkIdx >= 0; blkIdx--) {
+    if (state.tokens[blkIdx].type !== 'inline' ||
+        !QUOTE_TEST_RE.test(state.tokens[blkIdx].content)) {
+      continue
+    }
+
+    process_inlines(state.tokens[blkIdx].children, state)
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_core/state_core.mjs

@@ -0,0 +1,17 @@
+// Core state object
+//
+
+import Token from '../token.mjs'
+
+function StateCore (src, md, env) {
+  this.src = src
+  this.env = env
+  this.tokens = []
+  this.inlineMode = false
+  this.md = md // link to parser instance
+}
+
+// re-export Token class to use in core rules
+StateCore.prototype.Token = Token
+
+export default StateCore

mcp-server/node_modules/markdown-it/lib/rules_core/text_join.mjs

@@ -0,0 +1,43 @@
+// Join raw text tokens with the rest of the text
+//
+// This is set as a separate rule to provide an opportunity for plugins
+// to run text replacements after text join, but before escape join.
+//
+// For example, `\:)` shouldn't be replaced with an emoji.
+//
+
+export default function text_join (state) {
+  let curr, last
+  const blockTokens = state.tokens
+  const l = blockTokens.length
+
+  for (let j = 0; j < l; j++) {
+    if (blockTokens[j].type !== 'inline') continue
+
+    const tokens = blockTokens[j].children
+    const max = tokens.length
+
+    for (curr = 0; curr < max; curr++) {
+      if (tokens[curr].type === 'text_special') {
+        tokens[curr].type = 'text'
+      }
+    }
+
+    for (curr = last = 0; curr < max; curr++) {
+      if (tokens[curr].type === 'text' &&
+          curr + 1 < max &&
+          tokens[curr + 1].type === 'text') {
+        // collapse two adjacent text nodes
+        tokens[curr + 1].content = tokens[curr].content + tokens[curr + 1].content
+      } else {
+        if (curr !== last) { tokens[last] = tokens[curr] }
+
+        last++
+      }
+    }
+
+    if (curr !== last) {
+      tokens.length = last
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/autolink.mjs

@@ -0,0 +1,72 @@
+// Process autolinks '<protocol:...>'
+
+/* eslint max-len:0 */
+const EMAIL_RE    = /^([a-zA-Z0-9.!#$%&'*+/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*)$/
+/* eslint-disable-next-line no-control-regex */
+const AUTOLINK_RE = /^([a-zA-Z][a-zA-Z0-9+.-]{1,31}):([^<>\x00-\x20]*)$/
+
+export default function autolink (state, silent) {
+  let pos = state.pos
+
+  if (state.src.charCodeAt(pos) !== 0x3C/* < */) { return false }
+
+  const start = state.pos
+  const max = state.posMax
+
+  for (;;) {
+    if (++pos >= max) return false
+
+    const ch = state.src.charCodeAt(pos)
+
+    if (ch === 0x3C /* < */) return false
+    if (ch === 0x3E /* > */) break
+  }
+
+  const url = state.src.slice(start + 1, pos)
+
+  if (AUTOLINK_RE.test(url)) {
+    const fullUrl = state.md.normalizeLink(url)
+    if (!state.md.validateLink(fullUrl)) { return false }
+
+    if (!silent) {
+      const token_o   = state.push('link_open', 'a', 1)
+      token_o.attrs   = [['href', fullUrl]]
+      token_o.markup  = 'autolink'
+      token_o.info    = 'auto'
+
+      const token_t   = state.push('text', '', 0)
+      token_t.content = state.md.normalizeLinkText(url)
+
+      const token_c   = state.push('link_close', 'a', -1)
+      token_c.markup  = 'autolink'
+      token_c.info    = 'auto'
+    }
+
+    state.pos += url.length + 2
+    return true
+  }
+
+  if (EMAIL_RE.test(url)) {
+    const fullUrl = state.md.normalizeLink('mailto:' + url)
+    if (!state.md.validateLink(fullUrl)) { return false }
+
+    if (!silent) {
+      const token_o   = state.push('link_open', 'a', 1)
+      token_o.attrs   = [['href', fullUrl]]
+      token_o.markup  = 'autolink'
+      token_o.info    = 'auto'
+
+      const token_t   = state.push('text', '', 0)
+      token_t.content = state.md.normalizeLinkText(url)
+
+      const token_c   = state.push('link_close', 'a', -1)
+      token_c.markup  = 'autolink'
+      token_c.info    = 'auto'
+    }
+
+    state.pos += url.length + 2
+    return true
+  }
+
+  return false
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/backticks.mjs

@@ -0,0 +1,60 @@
+// Parse backticks
+
+export default function backtick (state, silent) {
+  let pos = state.pos
+  const ch = state.src.charCodeAt(pos)
+
+  if (ch !== 0x60/* ` */) { return false }
+
+  const start = pos
+  pos++
+  const max = state.posMax
+
+  // scan marker length
+  while (pos < max && state.src.charCodeAt(pos) === 0x60/* ` */) { pos++ }
+
+  const marker = state.src.slice(start, pos)
+  const openerLength = marker.length
+
+  if (state.backticksScanned && (state.backticks[openerLength] || 0) <= start) {
+    if (!silent) state.pending += marker
+    state.pos += openerLength
+    return true
+  }
+
+  let matchEnd = pos
+  let matchStart
+
+  // Nothing found in the cache, scan until the end of the line (or until marker is found)
+  while ((matchStart = state.src.indexOf('`', matchEnd)) !== -1) {
+    matchEnd = matchStart + 1
+
+    // scan marker length
+    while (matchEnd < max && state.src.charCodeAt(matchEnd) === 0x60/* ` */) { matchEnd++ }
+
+    const closerLength = matchEnd - matchStart
+
+    if (closerLength === openerLength) {
+      // Found matching closer length.
+      if (!silent) {
+        const token = state.push('code_inline', 'code', 0)
+        token.markup = marker
+        token.content = state.src.slice(pos, matchStart)
+          .replace(/\n/g, ' ')
+          .replace(/^ (.+) $/, '$1')
+      }
+      state.pos = matchEnd
+      return true
+    }
+
+    // Some different length found, put it in cache as upper limit of where closer can be found
+    state.backticks[closerLength] = matchStart
+  }
+
+  // Scanned through the end, didn't find anything
+  state.backticksScanned = true
+
+  if (!silent) state.pending += marker
+  state.pos += openerLength
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/balance_pairs.mjs

@@ -0,0 +1,124 @@
+// For each opening emphasis-like marker find a matching closing one
+//
+
+function processDelimiters (delimiters) {
+  const openersBottom = {}
+  const max = delimiters.length
+
+  if (!max) return
+
+  // headerIdx is the first delimiter of the current (where closer is) delimiter run
+  let headerIdx = 0
+  let lastTokenIdx = -2 // needs any value lower than -1
+  const jumps = []
+
+  for (let closerIdx = 0; closerIdx < max; closerIdx++) {
+    const closer = delimiters[closerIdx]
+
+    jumps.push(0)
+
+    // markers belong to same delimiter run if:
+    //  - they have adjacent tokens
+    //  - AND markers are the same
+    //
+    if (delimiters[headerIdx].marker !== closer.marker || lastTokenIdx !== closer.token - 1) {
+      headerIdx = closerIdx
+    }
+
+    lastTokenIdx = closer.token
+
+    // Length is only used for emphasis-specific "rule of 3",
+    // if it's not defined (in strikethrough or 3rd party plugins),
+    // we can default it to 0 to disable those checks.
+    //
+    closer.length = closer.length || 0
+
+    if (!closer.close) continue
+
+    // Previously calculated lower bounds (previous fails)
+    // for each marker, each delimiter length modulo 3,
+    // and for whether this closer can be an opener;
+    // https://github.com/commonmark/cmark/commit/34250e12ccebdc6372b8b49c44fab57c72443460
+    /* eslint-disable-next-line no-prototype-builtins */
+    if (!openersBottom.hasOwnProperty(closer.marker)) {
+      openersBottom[closer.marker] = [-1, -1, -1, -1, -1, -1]
+    }
+
+    const minOpenerIdx = openersBottom[closer.marker][(closer.open ? 3 : 0) + (closer.length % 3)]
+
+    let openerIdx = headerIdx - jumps[headerIdx] - 1
+
+    let newMinOpenerIdx = openerIdx
+
+    for (; openerIdx > minOpenerIdx; openerIdx -= jumps[openerIdx] + 1) {
+      const opener = delimiters[openerIdx]
+
+      if (opener.marker !== closer.marker) continue
+
+      if (opener.open && opener.end < 0) {
+        let isOddMatch = false
+
+        // from spec:
+        //
+        // If one of the delimiters can both open and close emphasis, then the
+        // sum of the lengths of the delimiter runs containing the opening and
+        // closing delimiters must not be a multiple of 3 unless both lengths
+        // are multiples of 3.
+        //
+        if (opener.close || closer.open) {
+          if ((opener.length + closer.length) % 3 === 0) {
+            if (opener.length % 3 !== 0 || closer.length % 3 !== 0) {
+              isOddMatch = true
+            }
+          }
+        }
+
+        if (!isOddMatch) {
+          // If previous delimiter cannot be an opener, we can safely skip
+          // the entire sequence in future checks. This is required to make
+          // sure algorithm has linear complexity (see *_*_*_*_*_... case).
+          //
+          const lastJump = openerIdx > 0 && !delimiters[openerIdx - 1].open
+            ? jumps[openerIdx - 1] + 1
+            : 0
+
+          jumps[closerIdx] = closerIdx - openerIdx + lastJump
+          jumps[openerIdx] = lastJump
+
+          closer.open  = false
+          opener.end   = closerIdx
+          opener.close = false
+          newMinOpenerIdx = -1
+          // treat next token as start of run,
+          // it optimizes skips in **<...>**a**<...>** pathological case
+          lastTokenIdx = -2
+          break
+        }
+      }
+    }
+
+    if (newMinOpenerIdx !== -1) {
+      // If match for this delimiter run failed, we want to set lower bound for
+      // future lookups. This is required to make sure algorithm has linear
+      // complexity.
+      //
+      // See details here:
+      // https://github.com/commonmark/cmark/issues/178#issuecomment-270417442
+      //
+      openersBottom[closer.marker][(closer.open ? 3 : 0) + ((closer.length || 0) % 3)] = newMinOpenerIdx
+    }
+  }
+}
+
+export default function link_pairs (state) {
+  const tokens_meta = state.tokens_meta
+  const max = state.tokens_meta.length
+
+  processDelimiters(state.delimiters)
+
+  for (let curr = 0; curr < max; curr++) {
+    if (tokens_meta[curr] && tokens_meta[curr].delimiters) {
+      processDelimiters(tokens_meta[curr].delimiters)
+    }
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/emphasis.mjs

@@ -0,0 +1,123 @@
+// Process *this* and _that_
+//
+
+// Insert each marker as a separate text token, and add it to delimiter list
+//
+function emphasis_tokenize (state, silent) {
+  const start = state.pos
+  const marker = state.src.charCodeAt(start)
+
+  if (silent) { return false }
+
+  if (marker !== 0x5F /* _ */ && marker !== 0x2A /* * */) { return false }
+
+  const scanned = state.scanDelims(state.pos, marker === 0x2A)
+
+  for (let i = 0; i < scanned.length; i++) {
+    const token = state.push('text', '', 0)
+    token.content = String.fromCharCode(marker)
+
+    state.delimiters.push({
+      // Char code of the starting marker (number).
+      //
+      marker,
+
+      // Total length of these series of delimiters.
+      //
+      length: scanned.length,
+
+      // A position of the token this delimiter corresponds to.
+      //
+      token: state.tokens.length - 1,
+
+      // If this delimiter is matched as a valid opener, `end` will be
+      // equal to its position, otherwise it's `-1`.
+      //
+      end: -1,
+
+      // Boolean flags that determine if this delimiter could open or close
+      // an emphasis.
+      //
+      open: scanned.can_open,
+      close: scanned.can_close
+    })
+  }
+
+  state.pos += scanned.length
+
+  return true
+}
+
+function postProcess (state, delimiters) {
+  const max = delimiters.length
+
+  for (let i = max - 1; i >= 0; i--) {
+    const startDelim = delimiters[i]
+
+    if (startDelim.marker !== 0x5F/* _ */ && startDelim.marker !== 0x2A/* * */) {
+      continue
+    }
+
+    // Process only opening markers
+    if (startDelim.end === -1) {
+      continue
+    }
+
+    const endDelim = delimiters[startDelim.end]
+
+    // If the previous delimiter has the same marker and is adjacent to this one,
+    // merge those into one strong delimiter.
+    //
+    // `<em><em>whatever</em></em>` -> `<strong>whatever</strong>`
+    //
+    const isStrong = i > 0 &&
+               delimiters[i - 1].end === startDelim.end + 1 &&
+               // check that first two markers match and adjacent
+               delimiters[i - 1].marker === startDelim.marker &&
+               delimiters[i - 1].token === startDelim.token - 1 &&
+               // check that last two markers are adjacent (we can safely assume they match)
+               delimiters[startDelim.end + 1].token === endDelim.token + 1
+
+    const ch = String.fromCharCode(startDelim.marker)
+
+    const token_o   = state.tokens[startDelim.token]
+    token_o.type    = isStrong ? 'strong_open' : 'em_open'
+    token_o.tag     = isStrong ? 'strong' : 'em'
+    token_o.nesting = 1
+    token_o.markup  = isStrong ? ch + ch : ch
+    token_o.content = ''
+
+    const token_c   = state.tokens[endDelim.token]
+    token_c.type    = isStrong ? 'strong_close' : 'em_close'
+    token_c.tag     = isStrong ? 'strong' : 'em'
+    token_c.nesting = -1
+    token_c.markup  = isStrong ? ch + ch : ch
+    token_c.content = ''
+
+    if (isStrong) {
+      state.tokens[delimiters[i - 1].token].content = ''
+      state.tokens[delimiters[startDelim.end + 1].token].content = ''
+      i--
+    }
+  }
+}
+
+// Walk through delimiter list and replace text tokens with tags
+//
+function emphasis_post_process (state) {
+  const tokens_meta = state.tokens_meta
+  const max = state.tokens_meta.length
+
+  postProcess(state, state.delimiters)
+
+  for (let curr = 0; curr < max; curr++) {
+    if (tokens_meta[curr] && tokens_meta[curr].delimiters) {
+      postProcess(state, tokens_meta[curr].delimiters)
+    }
+  }
+}
+
+export default {
+  tokenize: emphasis_tokenize,
+  postProcess: emphasis_post_process
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/entity.mjs

@@ -0,0 +1,51 @@
+// Process html entity - {, ¯, ", ...
+
+import { decodeHTML } from 'entities'
+import { isValidEntityCode, fromCodePoint } from '../common/utils.mjs'
+
+const DIGITAL_RE = /^&#((?:x[a-f0-9]{1,6}|[0-9]{1,7}));/i
+const NAMED_RE   = /^&([a-z][a-z0-9]{1,31});/i
+
+export default function entity (state, silent) {
+  const pos = state.pos
+  const max = state.posMax
+
+  if (state.src.charCodeAt(pos) !== 0x26/* & */) return false
+
+  if (pos + 1 >= max) return false
+
+  const ch = state.src.charCodeAt(pos + 1)
+
+  if (ch === 0x23 /* # */) {
+    const match = state.src.slice(pos).match(DIGITAL_RE)
+    if (match) {
+      if (!silent) {
+        const code = match[1][0].toLowerCase() === 'x' ? parseInt(match[1].slice(1), 16) : parseInt(match[1], 10)
+
+        const token   = state.push('text_special', '', 0)
+        token.content = isValidEntityCode(code) ? fromCodePoint(code) : fromCodePoint(0xFFFD)
+        token.markup  = match[0]
+        token.info    = 'entity'
+      }
+      state.pos += match[0].length
+      return true
+    }
+  } else {
+    const match = state.src.slice(pos).match(NAMED_RE)
+    if (match) {
+      const decoded = decodeHTML(match[0])
+      if (decoded !== match[0]) {
+        if (!silent) {
+          const token   = state.push('text_special', '', 0)
+          token.content = decoded
+          token.markup  = match[0]
+          token.info    = 'entity'
+        }
+        state.pos += match[0].length
+        return true
+      }
+    }
+  }
+
+  return false
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/escape.mjs

@@ -0,0 +1,69 @@
+// Process escaped chars and hardbreaks
+
+import { isSpace } from '../common/utils.mjs'
+
+const ESCAPED = []
+
+for (let i = 0; i < 256; i++) { ESCAPED.push(0) }
+
+'\\!"#$%&\'()*+,./:;<=>?@[]^_`{|}~-'
+  .split('').forEach(function (ch) { ESCAPED[ch.charCodeAt(0)] = 1 })
+
+export default function escape (state, silent) {
+  let pos = state.pos
+  const max = state.posMax
+
+  if (state.src.charCodeAt(pos) !== 0x5C/* \ */) return false
+  pos++
+
+  // '\' at the end of the inline block
+  if (pos >= max) return false
+
+  let ch1 = state.src.charCodeAt(pos)
+
+  if (ch1 === 0x0A) {
+    if (!silent) {
+      state.push('hardbreak', 'br', 0)
+    }
+
+    pos++
+    // skip leading whitespaces from next line
+    while (pos < max) {
+      ch1 = state.src.charCodeAt(pos)
+      if (!isSpace(ch1)) break
+      pos++
+    }
+
+    state.pos = pos
+    return true
+  }
+
+  let escapedStr = state.src[pos]
+
+  if (ch1 >= 0xD800 && ch1 <= 0xDBFF && pos + 1 < max) {
+    const ch2 = state.src.charCodeAt(pos + 1)
+
+    if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) {
+      escapedStr += state.src[pos + 1]
+      pos++
+    }
+  }
+
+  const origStr = '\\' + escapedStr
+
+  if (!silent) {
+    const token = state.push('text_special', '', 0)
+
+    if (ch1 < 256 && ESCAPED[ch1] !== 0) {
+      token.content = escapedStr
+    } else {
+      token.content = origStr
+    }
+
+    token.markup = origStr
+    token.info   = 'escape'
+  }
+
+  state.pos = pos + 1
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/fragments_join.mjs

@@ -0,0 +1,38 @@
+// Clean up tokens after emphasis and strikethrough postprocessing:
+// merge adjacent text nodes into one and re-calculate all token levels
+//
+// This is necessary because initially emphasis delimiter markers (*, _, ~)
+// are treated as their own separate text tokens. Then emphasis rule either
+// leaves them as text (needed to merge with adjacent text) or turns them
+// into opening/closing tags (which messes up levels inside).
+//
+
+export default function fragments_join (state) {
+  let curr, last
+  let level = 0
+  const tokens = state.tokens
+  const max = state.tokens.length
+
+  for (curr = last = 0; curr < max; curr++) {
+    // re-calculate levels after emphasis/strikethrough turns some text nodes
+    // into opening/closing tags
+    if (tokens[curr].nesting < 0) level-- // closing tag
+    tokens[curr].level = level
+    if (tokens[curr].nesting > 0) level++ // opening tag
+
+    if (tokens[curr].type === 'text' &&
+        curr + 1 < max &&
+        tokens[curr + 1].type === 'text') {
+      // collapse two adjacent text nodes
+      tokens[curr + 1].content = tokens[curr].content + tokens[curr + 1].content
+    } else {
+      if (curr !== last) { tokens[last] = tokens[curr] }
+
+      last++
+    }
+  }
+
+  if (curr !== last) {
+    tokens.length = last
+  }
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/html_inline.mjs

@@ -0,0 +1,50 @@
+// Process html tags
+
+import { HTML_TAG_RE } from '../common/html_re.mjs'
+
+function isLinkOpen (str) {
+  return /^<a[>\s]/i.test(str)
+}
+function isLinkClose (str) {
+  return /^<\/a\s*>/i.test(str)
+}
+
+function isLetter (ch) {
+  /* eslint no-bitwise:0 */
+  const lc = ch | 0x20 // to lower case
+  return (lc >= 0x61/* a */) && (lc <= 0x7a/* z */)
+}
+
+export default function html_inline (state, silent) {
+  if (!state.md.options.html) { return false }
+
+  // Check start
+  const max = state.posMax
+  const pos = state.pos
+  if (state.src.charCodeAt(pos) !== 0x3C/* < */ ||
+      pos + 2 >= max) {
+    return false
+  }
+
+  // Quick fail on second char
+  const ch = state.src.charCodeAt(pos + 1)
+  if (ch !== 0x21/* ! */ &&
+      ch !== 0x3F/* ? */ &&
+      ch !== 0x2F/* / */ &&
+      !isLetter(ch)) {
+    return false
+  }
+
+  const match = state.src.slice(pos).match(HTML_TAG_RE)
+  if (!match) { return false }
+
+  if (!silent) {
+    const token = state.push('html_inline', '', 0)
+    token.content = match[0]
+
+    if (isLinkOpen(token.content))  state.linkLevel++
+    if (isLinkClose(token.content)) state.linkLevel--
+  }
+  state.pos += match[0].length
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/image.mjs

@@ -0,0 +1,138 @@
+// Process ![image](<src> "title")
+
+import { normalizeReference, isSpace } from '../common/utils.mjs'
+
+export default function image (state, silent) {
+  let code, content, label, pos, ref, res, title, start
+  let href = ''
+  const oldPos = state.pos
+  const max = state.posMax
+
+  if (state.src.charCodeAt(state.pos) !== 0x21/* ! */) { return false }
+  if (state.src.charCodeAt(state.pos + 1) !== 0x5B/* [ */) { return false }
+
+  const labelStart = state.pos + 2
+  const labelEnd = state.md.helpers.parseLinkLabel(state, state.pos + 1, false)
+
+  // parser failed to find ']', so it's not a valid link
+  if (labelEnd < 0) { return false }
+
+  pos = labelEnd + 1
+  if (pos < max && state.src.charCodeAt(pos) === 0x28/* ( */) {
+    //
+    // Inline link
+    //
+
+    // [link](  <href>  "title"  )
+    //        ^^ skipping these spaces
+    pos++
+    for (; pos < max; pos++) {
+      code = state.src.charCodeAt(pos)
+      if (!isSpace(code) && code !== 0x0A) { break }
+    }
+    if (pos >= max) { return false }
+
+    // [link](  <href>  "title"  )
+    //          ^^^^^^ parsing link destination
+    start = pos
+    res = state.md.helpers.parseLinkDestination(state.src, pos, state.posMax)
+    if (res.ok) {
+      href = state.md.normalizeLink(res.str)
+      if (state.md.validateLink(href)) {
+        pos = res.pos
+      } else {
+        href = ''
+      }
+    }
+
+    // [link](  <href>  "title"  )
+    //                ^^ skipping these spaces
+    start = pos
+    for (; pos < max; pos++) {
+      code = state.src.charCodeAt(pos)
+      if (!isSpace(code) && code !== 0x0A) { break }
+    }
+
+    // [link](  <href>  "title"  )
+    //                  ^^^^^^^ parsing link title
+    res = state.md.helpers.parseLinkTitle(state.src, pos, state.posMax)
+    if (pos < max && start !== pos && res.ok) {
+      title = res.str
+      pos = res.pos
+
+      // [link](  <href>  "title"  )
+      //                         ^^ skipping these spaces
+      for (; pos < max; pos++) {
+        code = state.src.charCodeAt(pos)
+        if (!isSpace(code) && code !== 0x0A) { break }
+      }
+    } else {
+      title = ''
+    }
+
+    if (pos >= max || state.src.charCodeAt(pos) !== 0x29/* ) */) {
+      state.pos = oldPos
+      return false
+    }
+    pos++
+  } else {
+    //
+    // Link reference
+    //
+    if (typeof state.env.references === 'undefined') { return false }
+
+    if (pos < max && state.src.charCodeAt(pos) === 0x5B/* [ */) {
+      start = pos + 1
+      pos = state.md.helpers.parseLinkLabel(state, pos)
+      if (pos >= 0) {
+        label = state.src.slice(start, pos++)
+      } else {
+        pos = labelEnd + 1
+      }
+    } else {
+      pos = labelEnd + 1
+    }
+
+    // covers label === '' and label === undefined
+    // (collapsed reference link and shortcut reference link respectively)
+    if (!label) { label = state.src.slice(labelStart, labelEnd) }
+
+    ref = state.env.references[normalizeReference(label)]
+    if (!ref) {
+      state.pos = oldPos
+      return false
+    }
+    href = ref.href
+    title = ref.title
+  }
+
+  //
+  // We found the end of the link, and know for a fact it's a valid link;
+  // so all that's left to do is to call tokenizer.
+  //
+  if (!silent) {
+    content = state.src.slice(labelStart, labelEnd)
+
+    const tokens = []
+    state.md.inline.parse(
+      content,
+      state.md,
+      state.env,
+      tokens
+    )
+
+    const token = state.push('image', 'img', 0)
+    const attrs = [['src', href], ['alt', '']]
+    token.attrs = attrs
+    token.children = tokens
+    token.content = content
+
+    if (title) {
+      attrs.push(['title', title])
+    }
+  }
+
+  state.pos = pos
+  state.posMax = max
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/link.mjs

@@ -0,0 +1,139 @@
+// Process [link](<to> "stuff")
+
+import { normalizeReference, isSpace } from '../common/utils.mjs'
+
+export default function link (state, silent) {
+  let code, label, res, ref
+  let href = ''
+  let title = ''
+  let start = state.pos
+  let parseReference = true
+
+  if (state.src.charCodeAt(state.pos) !== 0x5B/* [ */) { return false }
+
+  const oldPos = state.pos
+  const max = state.posMax
+  const labelStart = state.pos + 1
+  const labelEnd = state.md.helpers.parseLinkLabel(state, state.pos, true)
+
+  // parser failed to find ']', so it's not a valid link
+  if (labelEnd < 0) { return false }
+
+  let pos = labelEnd + 1
+  if (pos < max && state.src.charCodeAt(pos) === 0x28/* ( */) {
+    //
+    // Inline link
+    //
+
+    // might have found a valid shortcut link, disable reference parsing
+    parseReference = false
+
+    // [link](  <href>  "title"  )
+    //        ^^ skipping these spaces
+    pos++
+    for (; pos < max; pos++) {
+      code = state.src.charCodeAt(pos)
+      if (!isSpace(code) && code !== 0x0A) { break }
+    }
+    if (pos >= max) { return false }
+
+    // [link](  <href>  "title"  )
+    //          ^^^^^^ parsing link destination
+    start = pos
+    res = state.md.helpers.parseLinkDestination(state.src, pos, state.posMax)
+    if (res.ok) {
+      href = state.md.normalizeLink(res.str)
+      if (state.md.validateLink(href)) {
+        pos = res.pos
+      } else {
+        href = ''
+      }
+
+      // [link](  <href>  "title"  )
+      //                ^^ skipping these spaces
+      start = pos
+      for (; pos < max; pos++) {
+        code = state.src.charCodeAt(pos)
+        if (!isSpace(code) && code !== 0x0A) { break }
+      }
+
+      // [link](  <href>  "title"  )
+      //                  ^^^^^^^ parsing link title
+      res = state.md.helpers.parseLinkTitle(state.src, pos, state.posMax)
+      if (pos < max && start !== pos && res.ok) {
+        title = res.str
+        pos = res.pos
+
+        // [link](  <href>  "title"  )
+        //                         ^^ skipping these spaces
+        for (; pos < max; pos++) {
+          code = state.src.charCodeAt(pos)
+          if (!isSpace(code) && code !== 0x0A) { break }
+        }
+      }
+    }
+
+    if (pos >= max || state.src.charCodeAt(pos) !== 0x29/* ) */) {
+      // parsing a valid shortcut link failed, fallback to reference
+      parseReference = true
+    }
+    pos++
+  }
+
+  if (parseReference) {
+    //
+    // Link reference
+    //
+    if (typeof state.env.references === 'undefined') { return false }
+
+    if (pos < max && state.src.charCodeAt(pos) === 0x5B/* [ */) {
+      start = pos + 1
+      pos = state.md.helpers.parseLinkLabel(state, pos)
+      if (pos >= 0) {
+        label = state.src.slice(start, pos++)
+      } else {
+        pos = labelEnd + 1
+      }
+    } else {
+      pos = labelEnd + 1
+    }
+
+    // covers label === '' and label === undefined
+    // (collapsed reference link and shortcut reference link respectively)
+    if (!label) { label = state.src.slice(labelStart, labelEnd) }
+
+    ref = state.env.references[normalizeReference(label)]
+    if (!ref) {
+      state.pos = oldPos
+      return false
+    }
+    href = ref.href
+    title = ref.title
+  }
+
+  //
+  // We found the end of the link, and know for a fact it's a valid link;
+  // so all that's left to do is to call tokenizer.
+  //
+  if (!silent) {
+    state.pos = labelStart
+    state.posMax = labelEnd
+
+    const token_o = state.push('link_open', 'a', 1)
+    const attrs = [['href', href]]
+    token_o.attrs  = attrs
+    if (title) {
+      attrs.push(['title', title])
+    }
+
+    state.linkLevel++
+    state.md.inline.tokenize(state)
+    state.linkLevel--
+
+    state.push('link_close', 'a', -1)
+  }
+
+  state.pos = pos
+  state.posMax = max
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/linkify.mjs

@@ -0,0 +1,56 @@
+// Process links like https://example.org/
+
+// RFC3986: scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
+const SCHEME_RE = /(?:^|[^a-z0-9.+-])([a-z][a-z0-9.+-]*)$/i
+
+export default function linkify (state, silent) {
+  if (!state.md.options.linkify) return false
+  if (state.linkLevel > 0) return false
+
+  const pos = state.pos
+  const max = state.posMax
+
+  if (pos + 3 > max) return false
+  if (state.src.charCodeAt(pos) !== 0x3A/* : */) return false
+  if (state.src.charCodeAt(pos + 1) !== 0x2F/* / */) return false
+  if (state.src.charCodeAt(pos + 2) !== 0x2F/* / */) return false
+
+  const match = state.pending.match(SCHEME_RE)
+  if (!match) return false
+
+  const proto = match[1]
+
+  const link = state.md.linkify.matchAtStart(state.src.slice(pos - proto.length))
+  if (!link) return false
+
+  let url = link.url
+
+  // invalid link, but still detected by linkify somehow;
+  // need to check to prevent infinite loop below
+  if (url.length <= proto.length) return false
+
+  // disallow '*' at the end of the link (conflicts with emphasis)
+  url = url.replace(/\*+$/, '')
+
+  const fullUrl = state.md.normalizeLink(url)
+  if (!state.md.validateLink(fullUrl)) return false
+
+  if (!silent) {
+    state.pending = state.pending.slice(0, -proto.length)
+
+    const token_o = state.push('link_open', 'a', 1)
+    token_o.attrs = [['href', fullUrl]]
+    token_o.markup = 'linkify'
+    token_o.info = 'auto'
+
+    const token_t = state.push('text', '', 0)
+    token_t.content = state.md.normalizeLinkText(url)
+
+    const token_c = state.push('link_close', 'a', -1)
+    token_c.markup = 'linkify'
+    token_c.info = 'auto'
+  }
+
+  state.pos += url.length - proto.length
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/newline.mjs

@@ -0,0 +1,42 @@
+// Proceess '\n'
+
+import { isSpace } from '../common/utils.mjs'
+
+export default function newline (state, silent) {
+  let pos = state.pos
+
+  if (state.src.charCodeAt(pos) !== 0x0A/* \n */) { return false }
+
+  const pmax = state.pending.length - 1
+  const max = state.posMax
+
+  // '  \n' -> hardbreak
+  // Lookup in pending chars is bad practice! Don't copy to other rules!
+  // Pending string is stored in concat mode, indexed lookups will cause
+  // convertion to flat mode.
+  if (!silent) {
+    if (pmax >= 0 && state.pending.charCodeAt(pmax) === 0x20) {
+      if (pmax >= 1 && state.pending.charCodeAt(pmax - 1) === 0x20) {
+        // Find whitespaces tail of pending chars.
+        let ws = pmax - 1
+        while (ws >= 1 && state.pending.charCodeAt(ws - 1) === 0x20) ws--
+
+        state.pending = state.pending.slice(0, ws)
+        state.push('hardbreak', 'br', 0)
+      } else {
+        state.pending = state.pending.slice(0, -1)
+        state.push('softbreak', 'br', 0)
+      }
+    } else {
+      state.push('softbreak', 'br', 0)
+    }
+  }
+
+  pos++
+
+  // skip heading spaces for next line
+  while (pos < max && isSpace(state.src.charCodeAt(pos))) { pos++ }
+
+  state.pos = pos
+  return true
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/state_inline.mjs

@@ -0,0 +1,123 @@
+// Inline parser state
+
+import Token from '../token.mjs'
+import { isWhiteSpace, isPunctChar, isMdAsciiPunct } from '../common/utils.mjs'
+
+function StateInline (src, md, env, outTokens) {
+  this.src = src
+  this.env = env
+  this.md = md
+  this.tokens = outTokens
+  this.tokens_meta = Array(outTokens.length)
+
+  this.pos = 0
+  this.posMax = this.src.length
+  this.level = 0
+  this.pending = ''
+  this.pendingLevel = 0
+
+  // Stores { start: end } pairs. Useful for backtrack
+  // optimization of pairs parse (emphasis, strikes).
+  this.cache = {}
+
+  // List of emphasis-like delimiters for current tag
+  this.delimiters = []
+
+  // Stack of delimiter lists for upper level tags
+  this._prev_delimiters = []
+
+  // backtick length => last seen position
+  this.backticks = {}
+  this.backticksScanned = false
+
+  // Counter used to disable inline linkify-it execution
+  // inside <a> and markdown links
+  this.linkLevel = 0
+}
+
+// Flush pending text
+//
+StateInline.prototype.pushPending = function () {
+  const token = new Token('text', '', 0)
+  token.content = this.pending
+  token.level = this.pendingLevel
+  this.tokens.push(token)
+  this.pending = ''
+  return token
+}
+
+// Push new token to "stream".
+// If pending text exists - flush it as text token
+//
+StateInline.prototype.push = function (type, tag, nesting) {
+  if (this.pending) {
+    this.pushPending()
+  }
+
+  const token = new Token(type, tag, nesting)
+  let token_meta = null
+
+  if (nesting < 0) {
+    // closing tag
+    this.level--
+    this.delimiters = this._prev_delimiters.pop()
+  }
+
+  token.level = this.level
+
+  if (nesting > 0) {
+    // opening tag
+    this.level++
+    this._prev_delimiters.push(this.delimiters)
+    this.delimiters = []
+    token_meta = { delimiters: this.delimiters }
+  }
+
+  this.pendingLevel = this.level
+  this.tokens.push(token)
+  this.tokens_meta.push(token_meta)
+  return token
+}
+
+// Scan a sequence of emphasis-like markers, and determine whether
+// it can start an emphasis sequence or end an emphasis sequence.
+//
+//  - start - position to scan from (it should point at a valid marker);
+//  - canSplitWord - determine if these markers can be found inside a word
+//
+StateInline.prototype.scanDelims = function (start, canSplitWord) {
+  const max = this.posMax
+  const marker = this.src.charCodeAt(start)
+
+  // treat beginning of the line as a whitespace
+  const lastChar = start > 0 ? this.src.charCodeAt(start - 1) : 0x20
+
+  let pos = start
+  while (pos < max && this.src.charCodeAt(pos) === marker) { pos++ }
+
+  const count = pos - start
+
+  // treat end of the line as a whitespace
+  const nextChar = pos < max ? this.src.charCodeAt(pos) : 0x20
+
+  const isLastPunctChar = isMdAsciiPunct(lastChar) || isPunctChar(String.fromCharCode(lastChar))
+  const isNextPunctChar = isMdAsciiPunct(nextChar) || isPunctChar(String.fromCharCode(nextChar))
+
+  const isLastWhiteSpace = isWhiteSpace(lastChar)
+  const isNextWhiteSpace = isWhiteSpace(nextChar)
+
+  const left_flanking =
+    !isNextWhiteSpace && (!isNextPunctChar || isLastWhiteSpace || isLastPunctChar)
+  const right_flanking =
+    !isLastWhiteSpace && (!isLastPunctChar || isNextWhiteSpace || isNextPunctChar)
+
+  const can_open  = left_flanking  && (canSplitWord || !right_flanking || isLastPunctChar)
+  const can_close = right_flanking && (canSplitWord || !left_flanking  || isNextPunctChar)
+
+  return { can_open, can_close, length: count }
+}
+
+// re-export Token class to use in block rules
+StateInline.prototype.Token = Token
+
+export default StateInline

mcp-server/node_modules/markdown-it/lib/rules_inline/strikethrough.mjs

@@ -0,0 +1,127 @@
+// ~~strike through~~
+//
+
+// Insert each marker as a separate text token, and add it to delimiter list
+//
+function strikethrough_tokenize (state, silent) {
+  const start = state.pos
+  const marker = state.src.charCodeAt(start)
+
+  if (silent) { return false }
+
+  if (marker !== 0x7E/* ~ */) { return false }
+
+  const scanned = state.scanDelims(state.pos, true)
+  let len = scanned.length
+  const ch = String.fromCharCode(marker)
+
+  if (len < 2) { return false }
+
+  let token
+
+  if (len % 2) {
+    token         = state.push('text', '', 0)
+    token.content = ch
+    len--
+  }
+
+  for (let i = 0; i < len; i += 2) {
+    token         = state.push('text', '', 0)
+    token.content = ch + ch
+
+    state.delimiters.push({
+      marker,
+      length: 0,     // disable "rule of 3" length checks meant for emphasis
+      token: state.tokens.length - 1,
+      end: -1,
+      open: scanned.can_open,
+      close: scanned.can_close
+    })
+  }
+
+  state.pos += scanned.length
+
+  return true
+}
+
+function postProcess (state, delimiters) {
+  let token
+  const loneMarkers = []
+  const max = delimiters.length
+
+  for (let i = 0; i < max; i++) {
+    const startDelim = delimiters[i]
+
+    if (startDelim.marker !== 0x7E/* ~ */) {
+      continue
+    }
+
+    if (startDelim.end === -1) {
+      continue
+    }
+
+    const endDelim = delimiters[startDelim.end]
+
+    token         = state.tokens[startDelim.token]
+    token.type    = 's_open'
+    token.tag     = 's'
+    token.nesting = 1
+    token.markup  = '~~'
+    token.content = ''
+
+    token         = state.tokens[endDelim.token]
+    token.type    = 's_close'
+    token.tag     = 's'
+    token.nesting = -1
+    token.markup  = '~~'
+    token.content = ''
+
+    if (state.tokens[endDelim.token - 1].type === 'text' &&
+        state.tokens[endDelim.token - 1].content === '~') {
+      loneMarkers.push(endDelim.token - 1)
+    }
+  }
+
+  // If a marker sequence has an odd number of characters, it's splitted
+  // like this: `~~~~~` -> `~` + `~~` + `~~`, leaving one marker at the
+  // start of the sequence.
+  //
+  // So, we have to move all those markers after subsequent s_close tags.
+  //
+  while (loneMarkers.length) {
+    const i = loneMarkers.pop()
+    let j = i + 1
+
+    while (j < state.tokens.length && state.tokens[j].type === 's_close') {
+      j++
+    }
+
+    j--
+
+    if (i !== j) {
+      token = state.tokens[j]
+      state.tokens[j] = state.tokens[i]
+      state.tokens[i] = token
+    }
+  }
+}
+
+// Walk through delimiter list and replace text tokens with tags
+//
+function strikethrough_postProcess (state) {
+  const tokens_meta = state.tokens_meta
+  const max = state.tokens_meta.length
+
+  postProcess(state, state.delimiters)
+
+  for (let curr = 0; curr < max; curr++) {
+    if (tokens_meta[curr] && tokens_meta[curr].delimiters) {
+      postProcess(state, tokens_meta[curr].delimiters)
+    }
+  }
+}
+
+export default {
+  tokenize: strikethrough_tokenize,
+  postProcess: strikethrough_postProcess
+}

mcp-server/node_modules/markdown-it/lib/rules_inline/text.mjs

@@ -0,0 +1,86 @@
+// Skip text characters for text token, place those to pending buffer
+// and increment current pos
+
+// Rule to skip pure text
+// '{}$%@~+=:' reserved for extentions
+
+// !, ", #, $, %, &, ', (, ), *, +, ,, -, ., /, :, ;, <, =, >, ?, @, [, \, ], ^, _, `, {, |, }, or ~
+
+// !!!! Don't confuse with "Markdown ASCII Punctuation" chars
+// http://spec.commonmark.org/0.15/#ascii-punctuation-character
+function isTerminatorChar (ch) {
+  switch (ch) {
+    case 0x0A/* \n */:
+    case 0x21/* ! */:
+    case 0x23/* # */:
+    case 0x24/* $ */:
+    case 0x25/* % */:
+    case 0x26/* & */:
+    case 0x2A/* * */:
+    case 0x2B/* + */:
+    case 0x2D/* - */:
+    case 0x3A/* : */:
+    case 0x3C/* < */:
+    case 0x3D/* = */:
+    case 0x3E/* > */:
+    case 0x40/* @ */:
+    case 0x5B/* [ */:
+    case 0x5C/* \ */:
+    case 0x5D/* ] */:
+    case 0x5E/* ^ */:
+    case 0x5F/* _ */:
+    case 0x60/* ` */:
+    case 0x7B/* { */:
+    case 0x7D/* } */:
+    case 0x7E/* ~ */:
+      return true
+    default:
+      return false
+  }
+}
+
+export default function text (state, silent) {
+  let pos = state.pos
+
+  while (pos < state.posMax && !isTerminatorChar(state.src.charCodeAt(pos))) {
+    pos++
+  }
+
+  if (pos === state.pos) { return false }
+
+  if (!silent) { state.pending += state.src.slice(state.pos, pos) }
+
+  state.pos = pos
+
+  return true
+}
+
+// Alternative implementation, for memory.
+//
+// It costs 10% of performance, but allows extend terminators list, if place it
+// to `ParserInline` property. Probably, will switch to it sometime, such
+// flexibility required.
+
+/*
+var TERMINATOR_RE = /[\n!#$%&*+\-:<=>@[\\\]^_`{}~]/;
+
+module.exports = function text(state, silent) {
+  var pos = state.pos,
+      idx = state.src.slice(pos).search(TERMINATOR_RE);
+
+  // first char is terminator -> empty text
+  if (idx === 0) { return false; }
+
+  // no terminator -> text till end of string
+  if (idx < 0) {
+    if (!silent) { state.pending += state.src.slice(pos); }
+    state.pos = state.src.length;
+    return true;
+  }
+
+  if (!silent) { state.pending += state.src.slice(pos, pos + idx); }
+
+  state.pos += idx;
+
+  return true;
+}; */

mcp-server/node_modules/markdown-it/lib/token.mjs

@@ -0,0 +1,191 @@
+// Token class
+
+/**
+ * class Token
+ **/
+
+/**
+ * new Token(type, tag, nesting)
+ *
+ * Create new token and fill passed properties.
+ **/
+function Token (type, tag, nesting) {
+  /**
+   * Token#type -> String
+   *
+   * Type of the token (string, e.g. "paragraph_open")
+   **/
+  this.type     = type
+
+  /**
+   * Token#tag -> String
+   *
+   * html tag name, e.g. "p"
+   **/
+  this.tag      = tag
+
+  /**
+   * Token#attrs -> Array
+   *
+   * Html attributes. Format: `[ [ name1, value1 ], [ name2, value2 ] ]`
+   **/
+  this.attrs    = null
+
+  /**
+   * Token#map -> Array
+   *
+   * Source map info. Format: `[ line_begin, line_end ]`
+   **/
+  this.map      = null
+
+  /**
+   * Token#nesting -> Number
+   *
+   * Level change (number in {-1, 0, 1} set), where:
+   *
+   * -  `1` means the tag is opening
+   * -  `0` means the tag is self-closing
+   * - `-1` means the tag is closing
+   **/
+  this.nesting  = nesting
+
+  /**
+   * Token#level -> Number
+   *
+   * nesting level, the same as `state.level`
+   **/
+  this.level    = 0
+
+  /**
+   * Token#children -> Array
+   *
+   * An array of child nodes (inline and img tokens)
+   **/
+  this.children = null
+
+  /**
+   * Token#content -> String
+   *
+   * In a case of self-closing tag (code, html, fence, etc.),
+   * it has contents of this tag.
+   **/
+  this.content  = ''
+
+  /**
+   * Token#markup -> String
+   *
+   * '*' or '_' for emphasis, fence string for fence, etc.
+   **/
+  this.markup   = ''
+
+  /**
+   * Token#info -> String
+   *
+   * Additional information:
+   *
+   * - Info string for "fence" tokens
+   * - The value "auto" for autolink "link_open" and "link_close" tokens
+   * - The string value of the item marker for ordered-list "list_item_open" tokens
+   **/
+  this.info     = ''
+
+  /**
+   * Token#meta -> Object
+   *
+   * A place for plugins to store an arbitrary data
+   **/
+  this.meta     = null
+
+  /**
+   * Token#block -> Boolean
+   *
+   * True for block-level tokens, false for inline tokens.
+   * Used in renderer to calculate line breaks
+   **/
+  this.block    = false
+
+  /**
+   * Token#hidden -> Boolean
+   *
+   * If it's true, ignore this element when rendering. Used for tight lists
+   * to hide paragraphs.
+   **/
+  this.hidden   = false
+}
+
+/**
+ * Token.attrIndex(name) -> Number
+ *
+ * Search attribute index by name.
+ **/
+Token.prototype.attrIndex = function attrIndex (name) {
+  if (!this.attrs) { return -1 }
+
+  const attrs = this.attrs
+
+  for (let i = 0, len = attrs.length; i < len; i++) {
+    if (attrs[i][0] === name) { return i }
+  }
+  return -1
+}
+
+/**
+ * Token.attrPush(attrData)
+ *
+ * Add `[ name, value ]` attribute to list. Init attrs if necessary
+ **/
+Token.prototype.attrPush = function attrPush (attrData) {
+  if (this.attrs) {
+    this.attrs.push(attrData)
+  } else {
+    this.attrs = [attrData]
+  }
+}
+
+/**
+ * Token.attrSet(name, value)
+ *
+ * Set `name` attribute to `value`. Override old value if exists.
+ **/
+Token.prototype.attrSet = function attrSet (name, value) {
+  const idx = this.attrIndex(name)
+  const attrData = [name, value]
+
+  if (idx < 0) {
+    this.attrPush(attrData)
+  } else {
+    this.attrs[idx] = attrData
+  }
+}
+
+/**
+ * Token.attrGet(name)
+ *
+ * Get the value of attribute `name`, or null if it does not exist.
+ **/
+Token.prototype.attrGet = function attrGet (name) {
+  const idx = this.attrIndex(name)
+  let value = null
+  if (idx >= 0) {
+    value = this.attrs[idx][1]
+  }
+  return value
+}
+
+/**
+ * Token.attrJoin(name, value)
+ *
+ * Join value to existing attribute via space. Or create new attribute if not
+ * exists. Useful to operate with token classes.
+ **/
+Token.prototype.attrJoin = function attrJoin (name, value) {
+  const idx = this.attrIndex(name)
+
+  if (idx < 0) {
+    this.attrPush([name, value])
+  } else {
+    this.attrs[idx][1] = this.attrs[idx][1] + ' ' + value
+  }
+}
+
+export default Token

mcp-server/node_modules/markdown-it/package.json

@@ -0,0 +1,92 @@
+{
+  "name": "markdown-it",
+  "version": "14.1.0",
+  "description": "Markdown-it - modern pluggable markdown parser.",
+  "keywords": [
+    "markdown",
+    "parser",
+    "commonmark",
+    "markdown-it",
+    "markdown-it-plugin"
+  ],
+  "repository": "markdown-it/markdown-it",
+  "license": "MIT",
+  "main": "dist/index.cjs.js",
+  "module": "index.mjs",
+  "exports": {
+    ".": {
+      "import": "./index.mjs",
+      "require": "./dist/index.cjs.js"
+    },
+    "./*": {
+      "require": "./*",
+      "import": "./*"
+    }
+  },
+  "bin": {
+    "markdown-it": "bin/markdown-it.mjs"
+  },
+  "scripts": {
+    "lint": "eslint .",
+    "test": "npm run lint && CJS_ONLY=1 npm run build && c8 --exclude dist --exclude test -r text -r html -r lcov mocha && node support/specsplit.mjs",
+    "doc": "node support/build_doc.mjs",
+    "gh-doc": "npm run doc && gh-pages -d apidoc -f",
+    "demo": "npm run lint && node support/build_demo.mjs",
+    "gh-demo": "npm run demo && gh-pages -d demo -f -b master -r git@github.com:markdown-it/markdown-it.github.io.git",
+    "build": "rollup -c support/rollup.config.mjs",
+    "benchmark-deps": "npm install --prefix benchmark/extra/ -g marked@0.3.6 commonmark@0.26.0 markdown-it/markdown-it.git#2.2.1",
+    "specsplit": "support/specsplit.mjs good -o test/fixtures/commonmark/good.txt && support/specsplit.mjs bad -o test/fixtures/commonmark/bad.txt && support/specsplit.mjs",
+    "todo": "grep 'TODO' -n -r ./lib 2>/dev/null",
+    "prepublishOnly": "npm test && npm run build && npm run gh-demo && npm run gh-doc"
+  },
+  "files": [
+    "index.mjs",
+    "lib/",
+    "dist/"
+  ],
+  "dependencies": {
+    "argparse": "^2.0.1",
+    "entities": "^4.4.0",
+    "linkify-it": "^5.0.0",
+    "mdurl": "^2.0.0",
+    "punycode.js": "^2.3.1",
+    "uc.micro": "^2.1.0"
+  },
+  "devDependencies": {
+    "@rollup/plugin-babel": "^6.0.4",
+    "@rollup/plugin-commonjs": "^25.0.7",
+    "@rollup/plugin-node-resolve": "^15.2.3",
+    "@rollup/plugin-terser": "^0.4.4",
+    "ansi": "^0.3.0",
+    "benchmark": "~2.1.0",
+    "c8": "^8.0.1",
+    "chai": "^4.2.0",
+    "eslint": "^8.4.1",
+    "eslint-config-standard": "^17.1.0",
+    "express": "^4.14.0",
+    "gh-pages": "^6.1.0",
+    "highlight.js": "^11.9.0",
+    "jest-worker": "^29.7.0",
+    "markdown-it-abbr": "^2.0.0",
+    "markdown-it-container": "^4.0.0",
+    "markdown-it-deflist": "^3.0.0",
+    "markdown-it-emoji": "^3.0.0",
+    "markdown-it-footnote": "^4.0.0",
+    "markdown-it-for-inline": "^2.0.1",
+    "markdown-it-ins": "^4.0.0",
+    "markdown-it-mark": "^4.0.0",
+    "markdown-it-sub": "^2.0.0",
+    "markdown-it-sup": "^2.0.0",
+    "markdown-it-testgen": "^0.1.3",
+    "mocha": "^10.2.0",
+    "ndoc": "^6.0.0",
+    "needle": "^3.0.0",
+    "rollup": "^4.5.0",
+    "shelljs": "^0.8.4",
+    "supertest": "^6.0.1"
+  },
+  "mocha": {
+    "inline-diffs": true,
+    "timeout": 60000
+  }
+}