Publish dsh-search-mcp fork for newer DSH and Bailian WebSearch MCP.

Based on gxpppp/dsh-search-mcp; includes DSH web settings fixes, Clash fake-IP URL policy, and Bailian default server patch.

Co-authored-by: Cursor <cursoragent@cursor.com>

test/runtime.test.js

@@ -0,0 +1,130 @@
+import test from 'node:test';
+import assert from 'node:assert/strict';
+import { clampSearchResults, SEARCH_MCP_CATALOG, resolveServer } from '../lib/catalog.js';
+import { extractSearchResult } from '../lib/extract.js';
+
+test('catalog exposes every supported provider preset', () => {
+  assert.deepEqual(Object.keys(SEARCH_MCP_CATALOG), [
+    'tavily',
+    'brave',
+    'exa',
+    'perplexity',
+    'duckduckgo',
+    'custom',
+  ]);
+});
+
+test('known providers ignore stored connection overrides', () => {
+  const tavily = resolveServer({
+    id: 'primary',
+    kind: 'tavily',
+    transport: 'stdio',
+    url: 'https://example.test/mcp',
+    authStyle: 'header',
+    authParam: 'X-Other-Key',
+    toolName: 'other_search',
+    apiKeyEnv: 'MY_TAVILY_KEY',
+    maxResults: 12,
+  });
+  assert.equal(tavily.transport, 'http');
+  assert.equal(tavily.url, 'https://mcp.tavily.com/mcp/');
+  assert.equal(tavily.authStyle, 'query');
+  assert.equal(tavily.authParam, 'tavilyApiKey');
+  assert.equal(tavily.toolName, 'tavily_search');
+  assert.equal(tavily.countArg, 'max_results');
+  assert.equal(tavily.apiKeyEnv, 'MY_TAVILY_KEY');
+  assert.equal(tavily.maxResults, 12);
+});
+
+test('prototype property kinds fall back to custom', () => {
+  for (const kind of ['constructor', 'toString', '__proto__']) {
+    const resolved = resolveServer({ id: kind, kind, url: 'https://search.example/mcp' });
+    assert.equal(resolved.kind, 'custom');
+    assert.equal(resolved.url, 'https://search.example/mcp');
+  }
+});
+
+test('custom providers preserve advanced connection fields', () => {
+  const custom = resolveServer({
+    id: 'custom',
+    kind: 'custom',
+    transport: 'stdio',
+    command: 'custom-mcp',
+    args: ['--stdio'],
+    authStyle: 'header',
+    authParam: 'X-Key',
+    authPrefix: 'Token ',
+    toolName: 'search',
+  });
+  assert.equal(custom.transport, 'stdio');
+  assert.equal(custom.command, 'custom-mcp');
+  assert.deepEqual(custom.args, ['--stdio']);
+  assert.equal(custom.authParam, 'X-Key');
+  assert.equal(custom.authPrefix, 'Token ');
+  assert.equal(custom.toolName, 'search');
+});
+
+test('provider contracts match current upstream transports', () => {
+  const brave = resolveServer({ id: 'brave', kind: 'brave' });
+  assert.equal(brave.transport, 'stdio');
+  assert.equal(brave.command, 'npx');
+  assert.deepEqual(brave.args, ['-y', '@brave/brave-search-mcp-server@2.1.3']);
+  assert.equal(brave.authParam, 'BRAVE_API_KEY');
+
+  const perplexity = resolveServer({ id: 'perplexity', kind: 'perplexity' });
+  assert.equal(perplexity.url, 'https://api.perplexity.ai/mcp');
+  assert.equal(perplexity.authParam, 'Authorization');
+  assert.equal(perplexity.authPrefix, 'Bearer ');
+  assert.equal(perplexity.toolName, 'perplexity_search');
+
+  const duckduckgo = resolveServer({ id: 'duckduckgo', kind: 'duckduckgo' });
+  assert.equal(duckduckgo.needsKey, false);
+  assert.equal(duckduckgo.toolName, 'duckduckgo_web_search');
+  assert.equal(duckduckgo.countArg, 'count');
+});
+
+test('provider result counts stay within upstream MCP schemas', () => {
+  assert.equal(clampSearchResults(resolveServer({ id: 't', kind: 'tavily' }), 1), 5);
+  assert.equal(clampSearchResults(resolveServer({ id: 't', kind: 'tavily' }), 50), 20);
+  assert.equal(clampSearchResults(resolveServer({ id: 'b', kind: 'brave' }), 50), 20);
+  assert.equal(clampSearchResults(resolveServer({ id: 'p', kind: 'perplexity' }), 0), 1);
+  assert.equal(clampSearchResults(resolveServer({ id: 'd', kind: 'duckduckgo' }), 50), 20);
+  assert.equal(clampSearchResults(resolveServer({ id: 'e', kind: 'exa' }), 50), 50);
+});
+test('extractSearchResult normalizes, deduplicates, and rejects invalid URLs', () => {
+  const result = extractSearchResult({
+    structuredContent: {
+      answer: 'Summary',
+      results: [
+        { url: 'https://example.com/a', title: 'A', content: 'alpha', published_date: '2026-08-18' },
+        { url: 'https://example.com/a', title: 'Duplicate' },
+        { url: 'ftp://example.com/ignored', title: 'Ignored' },
+      ],
+      nested: { url: 'http://example.com/b', description: 'beta' },
+    },
+  });
+
+  assert.deepEqual(result, {
+    sources: [
+      {
+        url: 'https://example.com/a',
+        title: 'A',
+        snippet: 'alpha',
+        publishedAt: '2026-08-18',
+      },
+      { url: 'http://example.com/b', snippet: 'beta' },
+    ],
+    truncated: false,
+    content: 'Summary',
+  });
+});
+
+test('extractSearchResult accepts JSON and plain text MCP blocks', () => {
+  const json = extractSearchResult({
+    content: [{ type: 'text', text: '{"results":[{"url":"https://example.com"}]}' }],
+  });
+  assert.equal(json.sources.length, 1);
+
+  const text = extractSearchResult({ content: [{ type: 'text', text: 'Direct answer' }] });
+  assert.deepEqual(text, { sources: [], truncated: false, content: 'Direct answer' });
+});