skannr 0.1.6 → 0.2.0

package/dist/languages/registry.d.ts

@@ -1,4 +1,10 @@
+/**
+ * Language adapter registry.
+ * Routes files to the appropriate parser based on extension.
+ */
 import { LanguageAdapter } from './LanguageAdapter';
 export declare function getAdapter(filePath: string): LanguageAdapter;
+/** Await full initialization of WASM-based adapters. */
+export declare function initializeAdapters(): Promise<void>;
 export declare function detectRepoLanguages(root: string): string[];
 //# sourceMappingURL=registry.d.ts.map

package/dist/languages/registry.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"registry.d.ts","sourceRoot":"","sources":["../../src/languages/registry.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,MAAM,mBAAmB,CAAC;AAapD,wBAAgB,UAAU,CAAC,QAAQ,EAAE,MAAM,GAAG,eAAe,CAG5D;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,CAQ1D"}
+{"version":3,"file":"registry.d.ts","sourceRoot":"","sources":["../../src/languages/registry.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,mBAAmB,CAAC;AAmBpD,wBAAgB,UAAU,CAAC,QAAQ,EAAE,MAAM,GAAG,eAAe,CAG5D;AAED,wDAAwD;AACxD,wBAAsB,kBAAkB,IAAI,OAAO,CAAC,IAAI,CAAC,CAExD;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,CAQ1D"}

package/dist/languages/registry.js

@@ -1,4 +1,8 @@
 "use strict";
+/**
+ * Language adapter registry.
+ * Routes files to the appropriate parser based on extension.
+ */
 var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
     if (k2 === undefined) k2 = k;
     var desc = Object.getOwnPropertyDescriptor(m, k);
@@ -34,20 +38,30 @@ var __importStar = (this && this.__importStar) || (function () {
 })();
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.getAdapter = getAdapter;
+exports.initializeAdapters = initializeAdapters;
 exports.detectRepoLanguages = detectRepoLanguages;
 const TypeScriptAdapter_1 = require("./TypeScriptAdapter");
-const PythonAdapter_1 = require("./PythonAdapter");
+const TreeSitterAdapter_1 = require("./TreeSitterAdapter");
 const GenericAdapter_1 = require("./GenericAdapter");
+const lang_config_1 = require("./lang-config");
 const path = __importStar(require("path"));
+const tsAdapter = new TypeScriptAdapter_1.TypeScriptAdapter();
+const treeSitterAdapter = new TreeSitterAdapter_1.TreeSitterAdapter(Object.values(lang_config_1.LANG_CONFIGS));
+const generic = new GenericAdapter_1.GenericAdapter();
+// Initialize tree-sitter WASM in background (non-blocking).
+treeSitterAdapter.initializeSync();
 const adapters = [
-    new TypeScriptAdapter_1.TypeScriptAdapter(),
-    new PythonAdapter_1.PythonAdapter(),
+    tsAdapter,
+    treeSitterAdapter,
 ];
-const generic = new GenericAdapter_1.GenericAdapter();
 function getAdapter(filePath) {
     const ext = path.extname(filePath).toLowerCase();
     return adapters.find((a) => a.extensions.includes(ext)) ?? generic;
 }
+/** Await full initialization of WASM-based adapters. */
+async function initializeAdapters() {
+    await treeSitterAdapter.initialize();
+}
 function detectRepoLanguages(root) {
     const fs = require('fs');
     const extCount = {};

package/dist/languages/registry.js.map

@@ -1 +1 @@
-{"version":3,"file":"registry.js","sourceRoot":"","sources":["../../src/languages/registry.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAaA,gCAGC;AAED,kDAQC;AAzBD,2DAAwD;AACxD,mDAAgD;AAChD,qDAAkD;AAClD,2CAA6B;AAE7B,MAAM,QAAQ,GAAsB;IAClC,IAAI,qCAAiB,EAAE;IACvB,IAAI,6BAAa,EAAE;CACpB,CAAC;AAEF,MAAM,OAAO,GAAG,IAAI,+BAAc,EAAE,CAAC;AAErC,SAAgB,UAAU,CAAC,QAAgB;IACzC,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IACjD,OAAO,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,OAAO,CAAC;AACrE,CAAC;AAED,SAAgB,mBAAmB,CAAC,IAAY;IAC9C,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IACzB,MAAM,QAAQ,GAA2B,EAAE,CAAC;IAC5C,YAAY,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;IACnC,OAAO,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC;SAC5B,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;SAC3B,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;SACX,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,SAAS,YAAY,CAAC,GAAW,EAAE,MAA8B,EAAE,KAAa,EAAE,QAAgB;IAChG,IAAI,KAAK,GAAG,QAAQ;QAAE,OAAO;IAC7B,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IACzB,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;IAC7B,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,cAAc,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC,CAAC;IAC5F,IAAI,CAAC;QACH,KAAK,MAAM,KAAK,IAAI,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;YACxC,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC;gBAAE,SAAS;YACjC,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;YACnC,MAAM,IAAI,GAAG,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;YAC/B,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;gBACvB,YAAY,CAAC,IAAI,EAAE,MAAM,EAAE,KAAK,GAAG,CAAC,EAAE,QAAQ,CAAC,CAAC;YAClD,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;gBAC9C,IAAI,GAAG;oBAAE,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;YAChD,CAAC;QACH,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,iCAAiC;IACnC,CAAC;AACH,CAAC"}
+{"version":3,"file":"registry.js","sourceRoot":"","sources":["../../src/languages/registry.ts"],"names":[],"mappings":";AAAA;;;GAGG;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAqBH,gCAGC;AAGD,gDAEC;AAED,kDAQC;AApCD,2DAAwD;AACxD,2DAAwD;AACxD,qDAAkD;AAClD,+CAA6C;AAC7C,2CAA6B;AAE7B,MAAM,SAAS,GAAG,IAAI,qCAAiB,EAAE,CAAC;AAC1C,MAAM,iBAAiB,GAAG,IAAI,qCAAiB,CAAC,MAAM,CAAC,MAAM,CAAC,0BAAY,CAAC,CAAC,CAAC;AAC7E,MAAM,OAAO,GAAG,IAAI,+BAAc,EAAE,CAAC;AAErC,4DAA4D;AAC5D,iBAAiB,CAAC,cAAc,EAAE,CAAC;AAEnC,MAAM,QAAQ,GAAsB;IAClC,SAAS;IACT,iBAAiB;CAClB,CAAC;AAEF,SAAgB,UAAU,CAAC,QAAgB;IACzC,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IACjD,OAAO,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,OAAO,CAAC;AACrE,CAAC;AAED,wDAAwD;AACjD,KAAK,UAAU,kBAAkB;IACtC,MAAM,iBAAiB,CAAC,UAAU,EAAE,CAAC;AACvC,CAAC;AAED,SAAgB,mBAAmB,CAAC,IAAY;IAC9C,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IACzB,MAAM,QAAQ,GAA2B,EAAE,CAAC;IAC5C,YAAY,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;IACnC,OAAO,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC;SAC5B,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;SAC3B,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;SACX,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,SAAS,YAAY,CAAC,GAAW,EAAE,MAA8B,EAAE,KAAa,EAAE,QAAgB;IAChG,IAAI,KAAK,GAAG,QAAQ;QAAE,OAAO;IAC7B,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IACzB,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;IAC7B,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,cAAc,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC,CAAC;IAC5F,IAAI,CAAC;QACH,KAAK,MAAM,KAAK,IAAI,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;YACxC,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC;gBAAE,SAAS;YACjC,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;YACnC,MAAM,IAAI,GAAG,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;YAC/B,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;gBACvB,YAAY,CAAC,IAAI,EAAE,MAAM,EAAE,KAAK,GAAG,CAAC,EAAE,QAAQ,CAAC,CAAC;YAClD,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;gBAC9C,IAAI,GAAG;oBAAE,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;YAChD,CAAC;QACH,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,iCAAiC;IACnC,CAAC;AACH,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "skannr",
-  "version": "0.1.6",
+  "version": "0.2.0",
   "description": "A universal tool to analyze any codebase and generate semantic skeletons for AI agents",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -11,7 +11,7 @@
     "skannr-mcp": "dist/mcp-cli.js"
   },
   "scripts": {
-    "build": "tsc",
+    "build": "tsc && node -e \"const fs=require('fs'),p=require('path');const s=p.join('src','languages','grammars'),d=p.join('dist','languages','grammars');fs.mkdirSync(d,{recursive:true});fs.readdirSync(s).filter(f=>f.endsWith('.wasm')).forEach(f=>fs.copyFileSync(p.join(s,f),p.join(d,f)));\"",
     "test": "jest",
     "lint": "eslint src --ext .ts",
     "start": "npm run build && node dist/cli.js",
@@ -34,11 +34,8 @@
     "commander": "^12.1.0",
     "ignore": "^7.0.5",
     "parse-diff": "0.12.0",
-    "ts-morph": "^21.0.1"
-  },
-  "optionalDependencies": {
-    "tree-sitter": "^0.25.0",
-    "tree-sitter-python": "^0.25.0"
+    "ts-morph": "^21.0.1",
+    "web-tree-sitter": "0.26.10"
   },
   "devDependencies": {
     "@types/jest": "^29.5.11",

package/src/languages/TreeSitterAdapter.ts

@@ -0,0 +1,431 @@
+/**
+ * Universal tree-sitter adapter using web-tree-sitter (WASM).
+ * Handles Python, Go, Rust, Java, and any future language by loading
+ * the appropriate grammar at runtime. No native compilation required.
+ */
+
+import * as path from 'path';
+import { LanguageAdapter, SkeletonResult, Symbol } from './LanguageAdapter';
+import { LangConfig, getLangConfigForExt } from './lang-config';
+
+// web-tree-sitter types (loaded dynamically to avoid ESM/CJS issues)
+let Parser: any = null;
+let initPromise: Promise<void> | null = null;
+const loadedLanguages = new Map<string, any>();
+
+/** Resolve path to a bundled .wasm grammar file. */
+function grammarPath(filename: string): string {
+  return path.join(__dirname, 'grammars', filename);
+}
+
+/** Initialize the WASM parser (once). */
+async function ensureInit(): Promise<boolean> {
+  if (Parser) return true;
+  if (initPromise) {
+    await initPromise;
+    return Parser !== null;
+  }
+  initPromise = (async () => {
+    try {
+      const mod = require('web-tree-sitter');
+      const P = mod.default ?? mod;
+      await P.init();
+      Parser = P;
+    } catch {
+      Parser = null;
+    }
+  })();
+  await initPromise;
+  return Parser !== null;
+}
+
+/** Load a language grammar (cached). */
+async function loadLanguage(config: LangConfig): Promise<any | null> {
+  if (loadedLanguages.has(config.id)) {
+    return loadedLanguages.get(config.id);
+  }
+  try {
+    const wasmPath = grammarPath(config.grammarFile);
+    const lang = await Parser.Language.load(wasmPath);
+    loadedLanguages.set(config.id, lang);
+    return lang;
+  } catch {
+    return null;
+  }
+}
+
+/** Synchronous wrapper — uses cached state if already initialized. */
+function getParserSync(config: LangConfig): { parser: any; language: any } | null {
+  if (!Parser) return null;
+  const lang = loadedLanguages.get(config.id);
+  if (!lang) return null;
+  const parser = new Parser();
+  parser.setLanguage(lang);
+  return { parser, language: lang };
+}
+
+/**
+ * Extract the text of a node's named field, or null if not present.
+ */
+function fieldText(node: any, fieldName: string): string | null {
+  const child = node.childForFieldName(fieldName);
+  return child ? child.text : null;
+}
+
+/**
+ * Get the signature of a function/method node (everything except the body).
+ */
+function getSignature(node: any, lines: string[], config: LangConfig): string {
+  const bodyNode = node.childForFieldName(config.bodyField);
+  if (!bodyNode) {
+    // No body field — return the whole node text (might be a declaration)
+    return node.text;
+  }
+  const startRow = node.startPosition.row;
+  const bodyStartRow = bodyNode.startPosition.row;
+  // Return lines from node start to body start (exclusive of body content)
+  const sigLines = lines.slice(startRow, bodyStartRow);
+  if (sigLines.length === 0) {
+    // Body is on the same line as the signature
+    const fullLine = lines[startRow] || '';
+    const bodyCol = bodyNode.startPosition.column;
+    return fullLine.slice(0, bodyCol).trimEnd();
+  }
+  return sigLines.join('\n');
+}
+
+/**
+ * Get the indent level of a node based on its start column.
+ */
+function indent(node: any): string {
+  return ' '.repeat(node.startPosition.column);
+}
+
+/**
+ * Determine the body placeholder based on language.
+ */
+function bodyPlaceholder(config: LangConfig): string {
+  if (config.id === 'python') return '    ...';
+  return '  /* trimmed */';
+}
+
+/**
+ * Generate skeleton for a class node.
+ */
+function skeletonizeClass(node: any, lines: string[], config: LangConfig): string[] {
+  const out: string[] = [];
+  const name = fieldText(node, config.nameField) || 'Unknown';
+  const bodyNode = node.childForFieldName(config.classBodyField);
+
+  // Class signature
+  const sig = getSignature(node, lines, config);
+  out.push(sig);
+
+  if (!bodyNode) {
+    return out;
+  }
+
+  // For Python, add body placeholder for the class opening
+  if (config.id === 'python') {
+    // Walk methods inside the class body
+    for (const child of bodyNode.namedChildren) {
+      if (config.methodTypes.includes(child.type)) {
+        out.push('');
+        const methodSig = getSignature(child, lines, config);
+        out.push(methodSig);
+        out.push(indent(child) + '    ...');
+      }
+    }
+  } else {
+    // For braced languages, output opening brace + method signatures
+    if (!sig.includes('{')) {
+      out.push(indent(node) + '{');
+    }
+    for (const child of bodyNode.namedChildren) {
+      if (config.methodTypes.includes(child.type)) {
+        const methodSig = getSignature(child, lines, config);
+        out.push(methodSig);
+        out.push(indent(child) + bodyPlaceholder(config));
+        if (!config.id.match(/python/)) {
+          out.push(indent(child) + '}');
+        }
+        out.push('');
+      } else if (child.type === 'field_declaration' ||
+                 child.type === 'field_definition' ||
+                 child.type === 'annotation_type_element_declaration') {
+        // Keep field declarations as-is
+        out.push(child.text);
+      }
+    }
+    out.push(indent(node) + '}');
+  }
+
+  return out;
+}
+
+/**
+ * Generate skeleton for a function node.
+ */
+function skeletonizeFunction(node: any, lines: string[], config: LangConfig): string[] {
+  const out: string[] = [];
+  const sig = getSignature(node, lines, config);
+  out.push(sig);
+
+  if (config.id === 'python') {
+    out.push(indent(node) + '    ...');
+  } else {
+    if (!sig.trimEnd().endsWith('{')) {
+      out.push(indent(node) + '{');
+    }
+    out.push(indent(node) + bodyPlaceholder(config));
+    out.push(indent(node) + '}');
+  }
+
+  return out;
+}
+
+export class TreeSitterAdapter implements LanguageAdapter {
+  name: string;
+  extensions: string[];
+  private configs: LangConfig[];
+  private _initDone = false;
+
+  constructor(configs: LangConfig[]) {
+    this.configs = configs;
+    this.name = 'tree-sitter';
+    this.extensions = configs.flatMap((c) => c.extensions);
+  }
+
+  canHandle(filePath: string): boolean {
+    const ext = path.extname(filePath).toLowerCase();
+    return this.extensions.includes(ext);
+  }
+
+  /** Ensure WASM parser is initialized. Call at the start of each public method. */
+  private ensureReady(ext: string): LangConfig | null {
+    const config = getLangConfigForExt(ext);
+    if (!config) return null;
+    if (!Parser || !loadedLanguages.has(config.id)) return null;
+    return config;
+  }
+
+  /** Async initialization — must be called once before using the adapter. */
+  async initialize(): Promise<void> {
+    if (this._initDone) return;
+    const ok = await ensureInit();
+    if (!ok) return;
+    for (const config of this.configs) {
+      await loadLanguage(config);
+    }
+    this._initDone = true;
+  }
+
+  /** Synchronous initialization attempt (for compat with existing sync API). */
+  initializeSync(): void {
+    // Trigger async init in background. Methods will return fallbacks
+    // until initialization completes.
+    if (!this._initDone) {
+      this.initialize().catch(() => {});
+    }
+  }
+
+  generateSkeleton(content: string, filePath: string): SkeletonResult {
+    const ext = path.extname(filePath).toLowerCase();
+    const config = this.ensureReady(ext);
+    if (!config) {
+      return { skeleton: content.split('\n').slice(0, 50).join('\n') };
+    }
+
+    try {
+      const ctx = getParserSync(config);
+      if (!ctx) return { skeleton: content.split('\n').slice(0, 50).join('\n') };
+
+      const tree = ctx.parser.parse(content);
+      if (!tree) return { skeleton: content.split('\n').slice(0, 50).join('\n') };
+
+      const lines = content.split(/\r?\n/);
+      const skeletonLines: string[] = [];
+
+      skeletonLines.push(`/* Skeleton of ${path.basename(filePath)} */`);
+      skeletonLines.push('');
+
+      for (const child of tree.rootNode.namedChildren) {
+        if (config.importTypes.includes(child.type)) {
+          skeletonLines.push(child.text);
+        } else if (config.classTypes.includes(child.type)) {
+          skeletonLines.push('');
+          skeletonLines.push(...skeletonizeClass(child, lines, config));
+        } else if (config.functionTypes.includes(child.type)) {
+          skeletonLines.push('');
+          skeletonLines.push(...skeletonizeFunction(child, lines, config));
+        } else if (config.interfaceTypes.includes(child.type)) {
+          skeletonLines.push('');
+          skeletonLines.push(child.text);
+        } else if (config.typeTypes.includes(child.type)) {
+          skeletonLines.push('');
+          skeletonLines.push(child.text);
+        }
+      }
+
+      tree.delete();
+      ctx.parser.delete();
+
+      return { skeleton: skeletonLines.join('\n').trim() };
+    } catch {
+      return { skeleton: content.split('\n').slice(0, 50).join('\n') };
+    }
+  }
+
+  extractSymbols(content: string): Symbol[] {
+    // Determine the config from context (we don't have filePath here,
+    // so try each config until one parses successfully)
+    for (const config of this.configs) {
+      const ctx = getParserSync(config);
+      if (!ctx) continue;
+
+      try {
+        const tree = ctx.parser.parse(content);
+        if (!tree) { ctx.parser.delete(); continue; }
+
+        const symbols: Symbol[] = [];
+        this.walkSymbols(tree.rootNode, symbols, config);
+        tree.delete();
+        ctx.parser.delete();
+
+        if (symbols.length > 0) return symbols;
+      } catch {
+        ctx.parser.delete();
+      }
+    }
+    return [];
+  }
+
+  extractImports(content: string): string[] {
+    for (const config of this.configs) {
+      const ctx = getParserSync(config);
+      if (!ctx) continue;
+
+      try {
+        const tree = ctx.parser.parse(content);
+        if (!tree) { ctx.parser.delete(); continue; }
+
+        const imports: string[] = [];
+        for (const child of tree.rootNode.namedChildren) {
+          if (config.importTypes.includes(child.type)) {
+            // Extract module name from import
+            const text = child.text;
+            imports.push(text);
+          }
+        }
+        tree.delete();
+        ctx.parser.delete();
+
+        if (imports.length > 0) return imports;
+      } catch {
+        ctx.parser.delete();
+      }
+    }
+    return [];
+  }
+
+  /** Extract symbols with a file path hint for correct config selection. */
+  extractSymbolsForFile(content: string, filePath: string): Symbol[] {
+    const ext = path.extname(filePath).toLowerCase();
+    const config = this.ensureReady(ext);
+    if (!config) return [];
+
+    const ctx = getParserSync(config);
+    if (!ctx) return [];
+
+    try {
+      const tree = ctx.parser.parse(content);
+      if (!tree) { ctx.parser.delete(); return []; }
+
+      const symbols: Symbol[] = [];
+      this.walkSymbols(tree.rootNode, symbols, config);
+      tree.delete();
+      ctx.parser.delete();
+      return symbols;
+    } catch {
+      ctx.parser.delete();
+      return [];
+    }
+  }
+
+  /** Extract imports with a file path hint for correct config selection. */
+  extractImportsForFile(content: string, filePath: string): string[] {
+    const ext = path.extname(filePath).toLowerCase();
+    const config = this.ensureReady(ext);
+    if (!config) return [];
+
+    const ctx = getParserSync(config);
+    if (!ctx) return [];
+
+    try {
+      const tree = ctx.parser.parse(content);
+      if (!tree) { ctx.parser.delete(); return []; }
+
+      const imports: string[] = [];
+      for (const child of tree.rootNode.namedChildren) {
+        if (config.importTypes.includes(child.type)) {
+          imports.push(child.text);
+        }
+      }
+      tree.delete();
+      ctx.parser.delete();
+      return imports;
+    } catch {
+      ctx.parser.delete();
+      return [];
+    }
+  }
+
+  private walkSymbols(node: any, symbols: Symbol[], config: LangConfig): void {
+    for (const child of node.namedChildren) {
+      const name = fieldText(child, config.nameField);
+      if (!name) continue;
+
+      if (config.functionTypes.includes(child.type)) {
+        symbols.push({
+          name,
+          kind: 'function',
+          line: child.startPosition.row + 1,
+        });
+      } else if (config.classTypes.includes(child.type)) {
+        symbols.push({
+          name,
+          kind: 'class',
+          line: child.startPosition.row + 1,
+        });
+        // Walk class body for methods
+        const body = child.childForFieldName(config.classBodyField);
+        if (body) {
+          for (const member of body.namedChildren) {
+            if (config.methodTypes.includes(member.type)) {
+              const methodName = fieldText(member, config.nameField);
+              if (methodName) {
+                symbols.push({
+                  name: methodName,
+                  kind: 'function',
+                  line: member.startPosition.row + 1,
+                });
+              }
+            }
+          }
+        }
+      } else if (config.interfaceTypes.includes(child.type)) {
+        symbols.push({
+          name,
+          kind: 'interface',
+          line: child.startPosition.row + 1,
+        });
+      } else if (config.typeTypes.includes(child.type)) {
+        symbols.push({
+          name,
+          kind: 'type',
+          line: child.startPosition.row + 1,
+        });
+      }
+    }
+  }
+}

package/src/languages/lang-config.ts

@@ -0,0 +1,113 @@
+/**
+ * Per-language AST node type mappings for tree-sitter grammars.
+ * Adding a new language = adding one entry to this map.
+ */
+
+export interface LangConfig {
+  /** Language identifier. */
+  id: string;
+  /** File extensions this language handles (lowercase, with dot). */
+  extensions: string[];
+  /** Filename of the .wasm grammar (relative to grammars/ dir). */
+  grammarFile: string;
+  /** AST node types that represent top-level function declarations. */
+  functionTypes: string[];
+  /** AST node types that represent class declarations. */
+  classTypes: string[];
+  /** AST node types that represent methods inside classes. */
+  methodTypes: string[];
+  /** AST node types that represent import statements. */
+  importTypes: string[];
+  /** AST node types that represent interface/protocol declarations. */
+  interfaceTypes: string[];
+  /** AST node types that represent type alias declarations. */
+  typeTypes: string[];
+  /** The field name used to extract the symbol name (usually 'name'). */
+  nameField: string;
+  /** The field name for the function/method body (usually 'body' or 'block'). */
+  bodyField: string;
+  /** The field name for class body. */
+  classBodyField: string;
+  /** The field name for function parameters. */
+  parametersField: string;
+  /** The field name for return type annotation (if the language supports it). */
+  returnTypeField: string;
+}
+
+export const LANG_CONFIGS: Record<string, LangConfig> = {
+  python: {
+    id: 'python',
+    extensions: ['.py', '.pyi'],
+    grammarFile: 'tree-sitter-python.wasm',
+    functionTypes: ['function_definition'],
+    classTypes: ['class_definition'],
+    methodTypes: ['function_definition'],
+    importTypes: ['import_statement', 'import_from_statement'],
+    interfaceTypes: [],
+    typeTypes: [],
+    nameField: 'name',
+    bodyField: 'body',
+    classBodyField: 'body',
+    parametersField: 'parameters',
+    returnTypeField: 'return_type',
+  },
+  go: {
+    id: 'go',
+    extensions: ['.go'],
+    grammarFile: 'tree-sitter-go.wasm',
+    functionTypes: ['function_declaration'],
+    classTypes: ['type_declaration'],
+    methodTypes: ['method_declaration'],
+    importTypes: ['import_declaration'],
+    interfaceTypes: ['type_declaration'],
+    typeTypes: ['type_declaration'],
+    nameField: 'name',
+    bodyField: 'body',
+    classBodyField: 'body',
+    parametersField: 'parameters',
+    returnTypeField: 'result',
+  },
+  rust: {
+    id: 'rust',
+    extensions: ['.rs'],
+    grammarFile: 'tree-sitter-rust.wasm',
+    functionTypes: ['function_item'],
+    classTypes: ['struct_item', 'enum_item'],
+    methodTypes: ['function_item'],
+    importTypes: ['use_declaration'],
+    interfaceTypes: ['trait_item'],
+    typeTypes: ['type_item'],
+    nameField: 'name',
+    bodyField: 'body',
+    classBodyField: 'body',
+    parametersField: 'parameters',
+    returnTypeField: 'return_type',
+  },
+  java: {
+    id: 'java',
+    extensions: ['.java'],
+    grammarFile: 'tree-sitter-java.wasm',
+    functionTypes: ['method_declaration', 'constructor_declaration'],
+    classTypes: ['class_declaration', 'enum_declaration'],
+    methodTypes: ['method_declaration', 'constructor_declaration'],
+    importTypes: ['import_declaration'],
+    interfaceTypes: ['interface_declaration'],
+    typeTypes: [],
+    nameField: 'name',
+    bodyField: 'body',
+    classBodyField: 'body',
+    parametersField: 'parameters',
+    returnTypeField: 'type',
+  },
+};
+
+/** Get the language config for a file extension, or null if unsupported. */
+export function getLangConfigForExt(ext: string): LangConfig | null {
+  const lower = ext.toLowerCase();
+  for (const config of Object.values(LANG_CONFIGS)) {
+    if (config.extensions.includes(lower)) {
+      return config;
+    }
+  }
+  return null;
+}