@gleanwork/mcp-server-tester 1.0.0-beta.7 → 1.0.0

package/README.md

@@ -126,7 +126,7 @@ LLM host mode makes real API calls and produces non-deterministic results. Use `
 Requires Node.js 22+.
 
 ```bash
-npm install --save-dev @gleanwork/mcp-server-tester @playwright/test zod
+npm install --save-dev @gleanwork/mcp-server-tester @playwright/test
 ```
 
 The Anthropic SDK is only needed for LLM-as-judge assertions or LLM host mode with the Anthropic provider:
@@ -182,6 +182,25 @@ For HTTP servers, set `transport: 'http'` and `serverUrl`. For servers that requ
 - [Development](./docs/development.md) — contributing and building
 - [Migration Guide (v0.12 → v1.0)](./docs/migrations/migration-1.0.md) — upgrading from pre-1.0 releases
 
+## AI Skills
+
+Install AI skills to help your coding assistant generate tests, eval datasets, and MCP host evals:
+
+```bash
+npx skills add -g gleanwork/mcp-server-tester
+```
+
+This installs skills globally so they're available across all your projects. Four skills are included:
+
+| Skill                 | Description                                                 |
+| --------------------- | ----------------------------------------------------------- |
+| `mcp-tester-guide`    | Framework reference — matchers, config, auth, anti-patterns |
+| `write-mcp-test`      | Generate direct-mode Playwright tests                       |
+| `write-mcp-eval`      | Generate data-driven eval datasets                          |
+| `write-mcp-host-eval` | Generate LLM host simulation evals                          |
+
+Compatible with Claude Code, Cursor, Windsurf, Copilot, and [40+ other AI agents](https://github.com/nicepkg/nice-skills).
+
 ## Examples
 
 The `examples/` directory contains complete working examples:

package/dist/cli/index.js

@@ -80,7 +80,7 @@ function JsonPreview({ data, maxLines = 15 }) {
 
 // package.json
 var package_default = {
-  version: "1.0.0-beta.7"};
+  version: "1.0.0"};
 
 // src/cli/templates/index.ts
 function getPlaywrightConfigTemplate(answers) {
@@ -1005,6 +1005,17 @@ async function createMCPClientForConfig(config, options) {
 }
 async function closeMCPClient(client) {
   try {
+    const transport = client.transport;
+    if (transport instanceof StreamableHTTPClientTransport) {
+      try {
+        await transport.terminateSession();
+      } catch (sessionError) {
+        debugClient(
+          "Error terminating session: %s",
+          sessionError instanceof Error ? sessionError.message : String(sessionError)
+        );
+      }
+    }
     await client.close();
   } catch (error) {
     debugClient(

package/dist/fixtures/mcp.d.ts

@@ -214,7 +214,7 @@ type RubricSpec = BuiltInRubric | {
 };
 
 /** Valid LLM judge provider kinds. */
-type ProviderKind = 'anthropic' | 'openai' | 'google';
+type ProviderKind = 'anthropic' | 'vertex-anthropic' | 'anthropic-agent-sdk' | 'openai' | 'google';
 
 /**
  * Tool call validators for mcp_host simulation results.
@@ -258,6 +258,12 @@ interface JudgeMatcherOptions {
     provider?: ProviderKind;
     /** Override the judge model */
     model?: string;
+    /**
+     * Name of a registered custom judge executor.
+     * When set, the named judge handles the entire evaluation pipeline
+     * and its `pass` result is authoritative.
+     */
+    judge?: string;
 }
 /**
  * Declaration merging for Playwright matchers
@@ -348,21 +354,30 @@ declare global {
              */
             toBeToolError(expected?: boolean | string | string[]): R;
             /**
-             * Validates that a response passes LLM-as-judge evaluation
+             * Validates that a response passes LLM-as-judge evaluation.
              *
-             * @param rubric - Evaluation rubric/criteria
-             * @param options - Judge options
+             * Two call signatures:
+             * - With rubric: `toPassToolJudge(rubric, options?)` — built-in LLM judge
+             * - With named judge: `toPassToolJudge({ judge: 'name' })` — custom judge executor
              *
              * @example
              * ```typescript
+             * // Built-in LLM judge with rubric
              * expect(result).toPassToolJudge('Response should be helpful and accurate');
-             * expect(result).toPassToolJudge('Response should match reference', {
+             * expect(result).toPassToolJudge('correctness', {
              *   reference: expectedOutput,
              *   passingThreshold: 0.8,
              * });
+             *
+             * // Named custom judge (registered via registerJudge)
+             * expect(result).toPassToolJudge({ judge: 'glean-completeness' });
              * ```
              */
             toPassToolJudge(rubric: RubricSpec, options?: JudgeMatcherOptions): Promise<R>;
+            toPassToolJudge(options: JudgeMatcherOptions): Promise<R>;
+            toPassToolJudge(judges: Array<JudgeMatcherOptions & {
+                rubric?: RubricSpec;
+            }>): Promise<R>;
             /**
              * Validates that a response meets size constraints
              *
@@ -452,16 +467,26 @@ type ToolPredicate = (response: unknown, text: string) => boolean | PredicateRes
  * Validates that a response passes LLM-as-judge evaluation.
  * Delegates evaluation logic to validateJudge() for consistency
  * with the validator/matcher duality pattern.
+ *
+ * Supports three call signatures:
+ *   - toPassToolJudge(rubric, options?)        — built-in LLM judge with rubric
+ *   - toPassToolJudge({ judge: 'name', ... })  — named custom judge
+ *   - toPassToolJudge([...judges])             — multi-judge (all must pass)
  */
 
 /**
- * Creates the toPassToolJudge matcher function
+ * The toPassToolJudge matcher function.
  *
- * Note: This is an async matcher that calls an LLM for evaluation.
+ * Accepts either:
+ *   (received, rubric, options?) — rubric-based LLM judge
+ *   (received, options)          — named custom judge (options.judge required)
+ *   (received, judges[])         — multi-judge (all must pass)
  */
 declare function toPassToolJudge(this: {
     isNot: boolean;
-}, received: unknown, rubric: RubricSpec, options?: JudgeMatcherOptions): Promise<{
+}, received: unknown, rubricOrOptions: RubricSpec | JudgeMatcherOptions | Array<JudgeMatcherOptions & {
+    rubric?: RubricSpec;
+}>, maybeOptions?: JudgeMatcherOptions): Promise<{
     pass: boolean;
     message: () => string;
 }>;