@pyreon/mcp 0.13.1 → 0.15.0

package/lib/types/index.d.ts

@@ -1 +1,7 @@
-export { };
+import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
+
+//#region src/index.d.ts
+declare function createServer(): McpServer;
+//#endregion
+export { createServer };
+//# sourceMappingURL=index2.d.ts.map

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pyreon/mcp",
-  "version": "0.13.1",
+  "version": "0.15.0",
   "description": "MCP server for Pyreon — AI-powered framework assistance",
   "homepage": "https://github.com/pyreon/pyreon/tree/main/packages/mcp#readme",
   "bugs": {
@@ -17,6 +17,7 @@
   },
   "files": [
     "lib",
+    "!lib/**/*.map",
     "src",
     "README.md",
     "LICENSE"
@@ -46,7 +47,7 @@
   },
   "dependencies": {
     "@modelcontextprotocol/sdk": "^1.29.0",
-    "@pyreon/compiler": "^0.13.1",
+    "@pyreon/compiler": "^0.15.0",
     "zod": "^4.3.6"
   },
   "devDependencies": {

package/src/anti-patterns.ts

@@ -0,0 +1,210 @@
+/**
+ * Parser for `.claude/rules/anti-patterns.md`. Drives the `get_anti_patterns`
+ * MCP tool (T2.5.4) and the `detector-tag-consistency` test, so a single
+ * canonical source of truth produces the AI-facing list, the doc file,
+ * and the drift guard.
+ *
+ * Format assumptions (enforced by the consistency test):
+ *  - Top-level category headings are `## <Name>` (second-level). The
+ *    first paragraph (before the first `## `) is intro prose and is
+ *    not returned as anti-patterns.
+ *  - Each anti-pattern is a line that starts with `- **Name**` at
+ *    column 0 and can continue onto subsequent lines (until the next
+ *    `- **` or `## `).
+ *  - An optional `[detector: <code>]` tag appears anywhere in the
+ *    bullet's first line — it pairs the bullet with a static
+ *    `PyreonDiagnosticCode`. Missing tag means the anti-pattern is
+ *    doc-only.
+ */
+
+export type AntiPatternCategory =
+  | 'reactivity'
+  | 'jsx'
+  | 'context'
+  | 'architecture'
+  | 'testing'
+  | 'lifecycle'
+  | 'documentation'
+
+export interface AntiPatternEntry {
+  /** Title extracted from `**...**` in the bullet */
+  name: string
+  /** Normalised category slug (matches the enum above) */
+  category: AntiPatternCategory
+  /** Category heading as it appears in the file (e.g. "Reactivity Mistakes") */
+  categoryHeading: string
+  /** Body text after the title, minus the detector tag */
+  description: string
+  /** Detector codes listed in `[detector: X / Y]` or `null` if none */
+  detectorCodes: string[]
+}
+
+// Heading → slug. Keep in sync with the anti-patterns.md section list.
+const CATEGORY_MAP: Record<string, AntiPatternCategory> = {
+  'Reactivity Mistakes': 'reactivity',
+  'JSX Mistakes': 'jsx',
+  'Context & Provider Mistakes': 'context',
+  'Architecture Mistakes': 'architecture',
+  'Testing Mistakes': 'testing',
+  'Lifecycle & Cleanup Mistakes': 'lifecycle',
+  'Documentation Mistakes': 'documentation',
+}
+
+export const ANTI_PATTERN_CATEGORIES: readonly AntiPatternCategory[] = [
+  'reactivity',
+  'jsx',
+  'context',
+  'architecture',
+  'testing',
+  'lifecycle',
+  'documentation',
+] as const
+
+function normaliseCategory(heading: string): AntiPatternCategory | null {
+  const trimmed = heading.trim()
+  return CATEGORY_MAP[trimmed] ?? null
+}
+
+function splitSections(doc: string): Array<{ heading: string; body: string }> {
+  const lines = doc.split('\n')
+  const sections: Array<{ heading: string; body: string }> = []
+  let currentHeading: string | null = null
+  let currentBody: string[] = []
+  for (const line of lines) {
+    const headingMatch = /^## (.+)$/.exec(line)
+    if (headingMatch) {
+      if (currentHeading !== null) {
+        sections.push({ heading: currentHeading, body: currentBody.join('\n') })
+      }
+      currentHeading = headingMatch[1]!
+      currentBody = []
+    } else if (currentHeading !== null) {
+      currentBody.push(line)
+    }
+  }
+  if (currentHeading !== null) {
+    sections.push({ heading: currentHeading, body: currentBody.join('\n') })
+  }
+  return sections
+}
+
+function splitBullets(sectionBody: string): string[] {
+  // Split on lines that start with `- **` at column 0. Continuation
+  // lines (any indented or non-bullet content) stay attached to the
+  // previous bullet.
+  const lines = sectionBody.split('\n')
+  const bullets: string[] = []
+  let current: string[] = []
+  for (const line of lines) {
+    if (/^- \*\*/.test(line)) {
+      if (current.length > 0) bullets.push(current.join('\n').trim())
+      current = [line]
+    } else if (current.length > 0) {
+      current.push(line)
+    }
+  }
+  if (current.length > 0) bullets.push(current.join('\n').trim())
+  return bullets.filter((b) => b.length > 0)
+}
+
+function parseBullet(bullet: string): {
+  name: string
+  description: string
+  detectorCodes: string[]
+} | null {
+  // `- **Name** [detector: ...]: body...` or `- **Name**: body...`
+  // Extract the **bolded** name first.
+  const nameMatch = /^- \*\*([^*]+)\*\*/.exec(bullet)
+  if (!nameMatch) return null
+  const name = nameMatch[1]!.trim()
+
+  const afterName = bullet.slice(nameMatch[0].length)
+
+  // Pull out the detector tag if present. It can appear as:
+  //   ` [detector: code]`
+  //   ` \`[detector: code]\``
+  const detectorMatch = /`?\[detector:\s*([a-z0-9\-\/ ]+)\]`?/i.exec(afterName)
+  const detectorCodes: string[] = []
+  if (detectorMatch) {
+    for (const code of detectorMatch[1]!.split('/')) {
+      const c = code.trim()
+      if (c) detectorCodes.push(c)
+    }
+  }
+
+  // Strip the detector tag + any leading `:` or spaces from the body.
+  let description = afterName
+  if (detectorMatch) {
+    description = description.replace(detectorMatch[0], '')
+  }
+  description = description.replace(/^[\s:]+/, '').trim()
+
+  return { name, description, detectorCodes }
+}
+
+export function parseAntiPatterns(doc: string): AntiPatternEntry[] {
+  const sections = splitSections(doc)
+  const entries: AntiPatternEntry[] = []
+  for (const { heading, body } of sections) {
+    const category = normaliseCategory(heading)
+    if (!category) continue
+    for (const bullet of splitBullets(body)) {
+      const parsed = parseBullet(bullet)
+      if (!parsed) continue
+      entries.push({
+        name: parsed.name,
+        category,
+        categoryHeading: heading,
+        description: parsed.description,
+        detectorCodes: parsed.detectorCodes,
+      })
+    }
+  }
+  return entries
+}
+
+/** Format a list of entries into a single Markdown block suitable for MCP. */
+export function formatAntiPatterns(
+  entries: AntiPatternEntry[],
+  filterCategory: AntiPatternCategory | 'all',
+): string {
+  if (entries.length === 0) {
+    return filterCategory === 'all'
+      ? 'No anti-patterns found. Check that `.claude/rules/anti-patterns.md` is reachable.'
+      : `No anti-patterns found in category '${filterCategory}'. Valid categories: ${ANTI_PATTERN_CATEGORIES.join(', ')}, all.`
+  }
+
+  // Group by category preserving the file order.
+  const byCategory = new Map<AntiPatternCategory, AntiPatternEntry[]>()
+  for (const entry of entries) {
+    if (!byCategory.has(entry.category)) byCategory.set(entry.category, [])
+    byCategory.get(entry.category)!.push(entry)
+  }
+
+  const parts: string[] = []
+  const header =
+    filterCategory === 'all'
+      ? `# Pyreon Anti-Patterns (${entries.length} total, ${byCategory.size} categor${byCategory.size === 1 ? 'y' : 'ies'})`
+      : `# Pyreon Anti-Patterns — ${filterCategory} (${entries.length})`
+  parts.push(header)
+  parts.push('')
+  parts.push(
+    'Each entry is a known mistake documented at `.claude/rules/anti-patterns.md`. Entries tagged `[detector: <code>]` are caught statically by the MCP `validate` tool — the rest require a human / AI review. Read them BEFORE writing new code, not during code review.',
+  )
+  parts.push('')
+
+  for (const [, catEntries] of byCategory) {
+    parts.push(`## ${catEntries[0]!.categoryHeading} (${catEntries.length})`)
+    parts.push('')
+    for (const entry of catEntries) {
+      const tag =
+        entry.detectorCodes.length > 0
+          ? ` \`[detector: ${entry.detectorCodes.join(' / ')}]\``
+          : ''
+      parts.push(`- **${entry.name}**${tag}: ${entry.description}`)
+    }
+    parts.push('')
+  }
+
+  return parts.join('\n').trimEnd()
+}