@mastra/mcp-docs-server 1.0.0-beta.9 → 1.0.0

package/.docs/raw/agents/guardrails.mdx

@@ -1,6 +1,8 @@
 ---
 title: "Guardrails | Agents"
 description: "Learn how to implement guardrails using input and output processors to secure and control AI interactions."
+packages:
+  - "@mastra/core"
 ---
 
 # Guardrails
@@ -22,7 +24,7 @@ Use processors for content moderation, prompt injection prevention, response san
 
 Import and instantiate the relevant processor class, and pass it to your agent’s configuration using either the `inputProcessors` or `outputProcessors` option:
 
-```typescript {2,8-16} title="src/mastra/agents/moderated-agent.ts" showLineNumbers copy
+```typescript {2,9-17} title="src/mastra/agents/moderated-agent.ts"
 import { Agent } from "@mastra/core/agent";
 import { ModerationProcessor } from "@mastra/core/processors";
 
@@ -51,13 +53,12 @@ Input processors are applied before user messages reach the language model. They
 
 The `UnicodeNormalizer` is an input processor that cleans and normalizes user input by unifying Unicode characters, standardizing whitespace, and removing problematic symbols, allowing the LLM to better understand user messages.
 
-```typescript {8-11} title="src/mastra/agents/normalized-agent.ts" showLineNumbers copy
+```typescript {7-10} title="src/mastra/agents/normalized-agent.ts"
 import { UnicodeNormalizer } from "@mastra/core/processors";
 
 export const normalizedAgent = new Agent({
   id: "normalized-agent",
   name: "Normalized Agent",
-  // ...
   inputProcessors: [
     new UnicodeNormalizer({
       stripControlChars: true,
@@ -67,19 +68,22 @@ export const normalizedAgent = new Agent({
 });
 ```
 
-> See [UnicodeNormalizer](/reference/v1/processors/unicode-normalizer) for a full list of configuration options.
+:::info
+
+Visit [UnicodeNormalizer](/reference/v1/processors/unicode-normalizer) for a full list of configuration options.
+
+:::
 
 ### Preventing prompt injection
 
 The `PromptInjectionDetector` is an input processor that scans user messages for prompt injection, jailbreak attempts, and system override patterns. It uses an LLM to classify risky input and can block or rewrite it before it reaches the model.
 
-```typescript {8-13} title="src/mastra/agents/secure-agent.ts" showLineNumbers copy
+```typescript {7-12} title="src/mastra/agents/secure-agent.ts"
 import { PromptInjectionDetector } from "@mastra/core/processors";
 
 export const secureAgent = new Agent({
   id: "secure-agent",
   name: "Secure Agent",
-  // ...
   inputProcessors: [
     new PromptInjectionDetector({
       model: "openrouter/openai/gpt-oss-safeguard-20b",
@@ -91,19 +95,22 @@ export const secureAgent = new Agent({
 });
 ```
 
-> See [PromptInjectionDetector](/reference/v1/processors/prompt-injection-detector) for a full list of configuration options.
+:::info
+
+Visit [PromptInjectionDetector](/reference/v1/processors/prompt-injection-detector) for a full list of configuration options.
+
+:::
 
 ### Detecting and translating language
 
 The `LanguageDetector` is an input processor that detects and translates user messages into a target language, enabling multilingual support while maintaining consistent interaction. It uses an LLM to identify the language and perform the translation.
 
-```typescript {8-13} title="src/mastra/agents/multilingual-agent.ts" showLineNumbers copy
+```typescript {7-12} title="src/mastra/agents/multilingual-agent.ts"
 import { LanguageDetector } from "@mastra/core/processors";
 
 export const multilingualAgent = new Agent({
   id: "multilingual-agent",
   name: "Multilingual Agent",
-  // ...
   inputProcessors: [
     new LanguageDetector({
       model: "openrouter/openai/gpt-oss-safeguard-20b",
@@ -115,7 +122,11 @@ export const multilingualAgent = new Agent({
 });
 ```
 
-> See [LanguageDetector](/reference/v1/processors/language-detector) for a full list of configuration options.
+:::info
+
+Visit [LanguageDetector](/reference/v1/processors/language-detector) for a full list of configuration options.
+
+:::
 
 ## Output processors
 
@@ -125,13 +136,12 @@ Output processors are applied after the language model generates a response, but
 
 The `BatchPartsProcessor` is an output processor that combines multiple stream parts before emitting them to the client. This reduces network overhead and improves the user experience by consolidating small chunks into larger batches.
 
-```typescript {8-12} title="src/mastra/agents/batched-agent.ts" showLineNumbers copy
+```typescript {7-11} title="src/mastra/agents/batched-agent.ts"
 import { BatchPartsProcessor } from "@mastra/core/processors";
 
 export const batchedAgent = new Agent({
   id: "batched-agent",
   name: "Batched Agent",
-  // ...
   outputProcessors: [
     new BatchPartsProcessor({
       batchSize: 5,
@@ -142,19 +152,22 @@ export const batchedAgent = new Agent({
 });
 ```
 
-> See [BatchPartsProcessor](/reference/v1/processors/batch-parts-processor) for a full list of configuration options.
+:::info
+
+Visit [BatchPartsProcessor](/reference/v1/processors/batch-parts-processor) for a full list of configuration options.
+
+:::
 
 ### Limiting token usage
 
 The `TokenLimiterProcessor` is an output processor that limits the number of tokens in model responses. It helps manage cost and performance by truncating or blocking messages when the limit is exceeded.
 
-```typescript {8-12} title="src/mastra/agents/limited-agent.ts" showLineNumbers copy
+```typescript {7-11} title="src/mastra/agents/limited-agent.ts"
 import { TokenLimiterProcessor } from "@mastra/core/processors";
 
 export const limitedAgent = new Agent({
   id: "limited-agent",
   name: "Limited Agent",
-  // ...
   outputProcessors: [
     new TokenLimiterProcessor({
       limit: 1000,
@@ -165,13 +178,17 @@ export const limitedAgent = new Agent({
 });
 ```
 
-> See [TokenLimiterProcessor](/reference/v1/processors/token-limiter-processor) for a full list of configuration options.
+:::info
+
+Visit [TokenLimiterProcessor](/reference/v1/processors/token-limiter-processor) for a full list of configuration options.
+
+:::
 
 ### Scrubbing system prompts
 
 The `SystemPromptScrubber` is an output processor that detects and redacts system prompts or other internal instructions from model responses. It helps prevent unintended disclosure of prompt content or configuration details that could introduce security risks. It uses an LLM to identify and redact sensitive content based on configured detection types.
 
-```typescript {7-16} title="src/mastra/agents/scrubbed-agent.ts" copy showLineNumbers
+```typescript {7-16} title="src/mastra/agents/scrubbed-agent.ts"
 import { SystemPromptScrubber } from "@mastra/core/processors";
 
 const scrubbedAgent = new Agent({
@@ -192,10 +209,14 @@ const scrubbedAgent = new Agent({
 });
 ```
 
-> See [SystemPromptScrubber](/reference/v1/processors/system-prompt-scrubber) for a full list of configuration options.
+:::info
+
+Visit [SystemPromptScrubber](/reference/v1/processors/system-prompt-scrubber) for a full list of configuration options.
+
+:::
 
 :::note
-When streaming responses over HTTP, Mastra redacts sensitive request data (system prompts, tool definitions, API keys) from stream chunks at the server level by default. See [Stream data redaction](/docs/v1/server-db/mastra-server#stream-data-redaction) for details.
+When streaming responses over HTTP, Mastra redacts sensitive request data (system prompts, tool definitions, API keys) from stream chunks at the server level by default. See [Stream data redaction](/docs/v1/server/mastra-server#stream-data-redaction) for details.
 :::
 
 ## Hybrid processors
@@ -206,13 +227,12 @@ Hybrid processors can be applied either before messages are sent to the language
 
 The `ModerationProcessor` is a hybrid processor that detects inappropriate or harmful content across categories like hate, harassment, and violence. It can be used to moderate either user input or model output, depending on where it's applied. It uses an LLM to classify the message and can block or rewrite it based on your configuration.
 
-```typescript {8-13,16-18} title="src/mastra/agents/moderated-agent.ts" showLineNumbers copy
+```typescript {7-12,15} title="src/mastra/agents/moderated-agent.ts"
 import { ModerationProcessor } from "@mastra/core/processors";
 
 export const moderatedAgent = new Agent({
   id: "moderated-agent",
   name: "Moderated Agent",
-  // ...
   inputProcessors: [
     new ModerationProcessor({
       model: "openrouter/openai/gpt-oss-safeguard-20b",
@@ -222,26 +242,27 @@ export const moderatedAgent = new Agent({
     }),
   ],
   outputProcessors: [
-    new ModerationProcessor({
-      // ...
-    }),
+    new ModerationProcessor(),
   ],
 });
 ```
 
-> See [ModerationProcessor](/reference/v1/processors/moderation-processor) for a full list of configuration options.
+:::info
+
+Visit [ModerationProcessor](/reference/v1/processors/moderation-processor) for a full list of configuration options.
+
+:::
 
 ### Detecting and redacting PII
 
 The `PIIDetector` is a hybrid processor that detects and removes personally identifiable information such as emails, phone numbers, and credit cards. It can redact either user input or model output, depending on where it's applied. It uses an LLM to identify sensitive content based on configured detection types.
 
-```typescript {8-15,18-20} title="src/mastra/agents/private-agent.ts" showLineNumbers copy
+```typescript {7-14,17} title="src/mastra/agents/private-agent.ts"
 import { PIIDetector } from "@mastra/core/processors";
 
 export const privateAgent = new Agent({
   id: "private-agent",
   name: "Private Agent",
-  // ...
   inputProcessors: [
     new PIIDetector({
       model: "openrouter/openai/gpt-oss-safeguard-20b",
@@ -253,14 +274,16 @@ export const privateAgent = new Agent({
     }),
   ],
   outputProcessors: [
-    new PIIDetector({
-      // ...
-    }),
+    new PIIDetector(),
   ],
 });
 ```
 
-> See [PIIDetector](/reference/v1/processors/pii-detector) for a full list of configuration options.
+:::info
+
+Visit [PIIDetector](/reference/v1/processors/pii-detector) for a full list of configuration options.
+
+:::
 
 ## Applying multiple processors
 
@@ -274,7 +297,7 @@ A typical order might be:
 
 The order affects behavior, so arrange processors to suit your goals.
 
-```typescript title="src/mastra/agents/test-agent.ts" showLineNumbers copy
+```typescript title="src/mastra/agents/test-agent.ts"
 import {
   UnicodeNormalizer,
   ModerationProcessor,
@@ -285,20 +308,11 @@ import {
 export const testAgent = new Agent({
   id: "test-agent",
   name: "Test Agent",
-  // ...
   inputProcessors: [
-    new UnicodeNormalizer({
-      //...
-    }),
-    new PromptInjectionDetector({
-      // ...
-    }),
-    new PIIDetector({
-      // ...
-    }),
-    new ModerationProcessor({
-      // ...
-    }),
+    new UnicodeNormalizer(),
+    new PromptInjectionDetector(),
+    new PIIDetector(),
+    new ModerationProcessor(),
   ],
 });
 ```
@@ -309,16 +323,14 @@ Many of the built-in processors support a `strategy` parameter that controls how
 
 Most strategies allow the request to continue without interruption. When `block` is used, the processor calls its internal `abort()` function, which immediately stops the request and prevents any subsequent processors from running.
 
-```typescript {10} title="src/mastra/agents/private-agent.ts" showLineNumbers copy
+```typescript {8} title="src/mastra/agents/private-agent.ts"
 import { PIIDetector } from "@mastra/core/processors";
 
 export const privateAgent = new Agent({
   id: "private-agent",
   name: "Private Agent",
-  // ...
   inputProcessors: [
     new PIIDetector({
-      // ...
       strategy: "block",
     }),
   ],
@@ -327,41 +339,78 @@ export const privateAgent = new Agent({
 
 ### Handling blocked requests
 
-When a processor blocks a request, the agent will still return successfully without throwing an error. To handle blocked requests, check for `tripwire` or `tripwireReason` in the response.
+When a processor blocks a request, the agent will still return successfully without throwing an error. To handle blocked requests, check for `tripwire` in the response.
 
-For example, if an agent uses the `PIIDetector` with `strategy: "block"` and the request includes a credit card number, it will be blocked and the response will include a `tripwireReason`.
+For example, if an agent uses the `PIIDetector` with `strategy: "block"` and the request includes a credit card number, it will be blocked and the response will include tripwire information.
 
 #### `.generate()` example
 
-```typescript showLineNumbers
+```typescript
 const result = await agent.generate(
   "Is this credit card number valid?: 4543 1374 5089 4332",
 );
 
-console.error(result.tripwire);
-console.error(result.tripwireReason);
+if (result.tripwire) {
+  console.error("Blocked:", result.tripwire.reason);
+  console.error("Processor:", result.tripwire.processorId);
+  // Optional: check if retry was requested
+  console.error("Retry requested:", result.tripwire.retry);
+  // Optional: access additional metadata
+  console.error("Metadata:", result.tripwire.metadata);
+}
 ```
 
 #### `.stream()` example
 
-```typescript showLineNumbers
+```typescript
 const stream = await agent.stream(
   "Is this credit card number valid?: 4543 1374 5089 4332",
 );
 
 for await (const chunk of stream.fullStream) {
   if (chunk.type === "tripwire") {
-    console.error(chunk.payload.tripwireReason);
+    console.error("Blocked:", chunk.payload.reason);
+    console.error("Processor:", chunk.payload.processorId);
   }
 }
 ```
 
-In this case, the `tripwireReason` indicates that a credit card number was detected:
+In this case, the `reason` indicates that a credit card number was detected:
 
 ```text
 PII detected. Types: credit-card
 ```
 
+### Requesting retries
+
+Processors can request that the LLM retry its response with feedback. This is useful for implementing quality checks:
+
+```typescript
+export class QualityChecker implements Processor {
+  id = "quality-checker";
+
+  async processOutputStep({ text, abort, retryCount }) {
+    const score = await evaluateQuality(text);
+
+    if (score < 0.7 && retryCount < 3) {
+      // Request retry with feedback for the LLM
+      abort("Response quality too low. Please be more specific.", {
+        retry: true,
+        metadata: { score },
+      });
+    }
+
+    return [];
+  }
+}
+```
+
+The `abort()` function accepts an optional second parameter with:
+- `retry: true` - Request the LLM retry the step
+- `metadata: unknown` - Attach additional data for debugging/logging
+
+Use `retryCount` to track retry attempts and prevent infinite loops.
+
 ## Custom processors
 
 If the built-in processors don’t cover your needs, you can create your own by extending the `Processor` class.

package/.docs/raw/agents/network-approval.mdx

@@ -0,0 +1,281 @@
+---
+title: "Network Approval | Agents"
+description: Learn how to require approvals, suspend execution, and resume suspended networks while keeping humans in control of agent network workflows.
+packages:
+  - "@mastra/core"
+  - "@mastra/libsql"
+  - "@mastra/memory"
+---
+
+# Network Approval
+
+Agent networks can require the same [human-in-the-loop](/docs/v1/workflows/human-in-the-loop) oversight used in individual agents and workflows. When a tool, sub-agent, or workflow within a network requires approval or suspends execution, the network pauses and emits events that allow your application to collect user input before resuming.
+
+## Storage
+
+Network approval uses snapshots to capture execution state. Ensure you've enabled a storage provider in your Mastra instance. If storage isn't enabled you'll see an error relating to snapshot not found.
+
+```typescript title="src/mastra/index.ts"
+import { Mastra } from "@mastra/core/mastra";
+import { LibSQLStore } from "@mastra/libsql";
+
+export const mastra = new Mastra({
+  storage: new LibSQLStore({
+    id: "mastra-storage",
+    url: ":memory:"
+  })
+});
+```
+
+## Approving network tool calls
+
+When a tool within a network has `requireApproval: true`, the network stream emits an `agent-execution-approval` chunk and pauses. To allow the tool to execute, call `approveNetworkToolCall` with the `runId`.
+
+```typescript
+const stream = await routingAgent.network("Process this query", {
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+
+let runId: string;
+
+for await (const chunk of stream) {
+  runId = stream.runId;
+  // if the requirApproval is in a tool inside a subAgent or the subAgent has requireToolApproval set to true
+  if (chunk.type === "agent-execution-approval") { 
+    console.log("Tool requires approval:", chunk.payload);
+  }
+
+  // if the requirApproval is in a tool directly in the network agent
+  if (chunk.type === "tool-execution-approval") {
+    console.log("Tool requires approval:", chunk.payload);
+  }
+}
+
+// Approve and resume execution
+const approvedStream = await routingAgent.approveNetworkToolCall({
+  runId,
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+
+for await (const chunk of approvedStream) {
+  if (chunk.type === "network-execution-event-step-finish") {
+    console.log(chunk.payload.result);
+  }
+}
+```
+
+## Declining network tool calls
+
+To decline a pending tool call and prevent execution, call `declineNetworkToolCall`. The network continues without executing the tool.
+
+```typescript
+const declinedStream = await routingAgent.declineNetworkToolCall({
+  runId,
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+
+for await (const chunk of declinedStream) {
+  if (chunk.type === "network-execution-event-step-finish") {
+    console.log(chunk.payload.result);
+  }
+}
+```
+
+## Resuming suspended networks
+
+When a primitive in the network calls `suspend()`, the stream emits an `agent-execution-suspended`/`tool-execution-suspended`/`workflow-execution-suspended` chunk with a `suspendPayload` containing context from the primitive. Use `resumeNetwork` to provide the data requested by the primitive and continue execution.
+
+```typescript
+import { createTool } from "@mastra/core/tools";
+import { z } from "zod";
+
+const confirmationTool = createTool({
+  id: "confirmation-tool",
+  description: "Requests user confirmation before proceeding",
+  inputSchema: z.object({
+    action: z.string()
+  }),
+  outputSchema: z.object({
+    confirmed: z.boolean(),
+    action: z.string()
+  }),
+  suspendSchema: z.object({
+    message: z.string(),
+    action: z.string()
+  }),
+  resumeSchema: z.object({
+    confirmed: z.boolean()
+  }),
+  execute: async (inputData, context) => {
+    const { resumeData, suspend } = context?.agent ?? {};
+
+    if (!resumeData?.confirmed) {
+      return suspend?.({
+        message: `Please confirm: ${inputData.action}`,
+        action: inputData.action
+      });
+    }
+
+    return { confirmed: true, action: inputData.action };
+  }
+});
+```
+
+Handle the suspension and resume with user-provided data:
+
+```typescript
+const stream = await routingAgent.network("Delete the old records", {
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+
+for await (const chunk of stream) {
+  if (chunk.type === "workflow-execution-suspended") {
+    console.log(chunk.payload.suspendPayload);
+    // { message: "Please confirm: delete old records", action: "delete old records" }
+  }
+}
+
+// Resume with user confirmation
+const resumedStream = await routingAgent.resumeNetwork(
+  { confirmed: true },
+  {
+    runId: stream.runId,
+    memory: {
+      thread: "user-123",
+      resource: "my-app"
+    }
+  }
+);
+
+for await (const chunk of resumedStream) {
+  if (chunk.type === "network-execution-event-step-finish") {
+    console.log(chunk.payload.result);
+  }
+}
+```
+
+## Automatic primitive resumption
+
+When using primitives that call `suspend()`, you can enable automatic resumption so the network resumes suspended primitives based on the user's next message. This creates a conversational flow where users provide the required information naturally.
+
+### Enabling auto-resume
+
+Set `autoResumeSuspendedTools` to `true` in the agent's `defaultNetworkOptions` or when calling `network()`:
+
+```typescript
+import { Agent } from "@mastra/core/agent";
+import { Memory } from "@mastra/memory";
+
+// Option 1: In agent configuration
+const routingAgent = new Agent({
+  id: "routing-agent",
+  name: "Routing Agent",
+  instructions: "You coordinate tasks across multiple agents",
+  model: "openai/gpt-4o-mini",
+  tools: { confirmationTool },
+  memory: new Memory(),
+  defaultNetworkOptions: {
+    autoResumeSuspendedTools: true,
+  },
+});
+
+// Option 2: Per-request
+const stream = await routingAgent.network("Process this request", {
+  autoResumeSuspendedTools: true,
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+```
+
+### How it works
+
+When `autoResumeSuspendedTools` is enabled:
+
+1. A primitive suspends execution by calling `suspend()` with a payload
+2. The suspension is persisted to memory along with the conversation
+3. When the user sends their next message on the same thread, the network:
+   - Detects the suspended primitive from message history
+   - Extracts `resumeData` from the user's message based on the tool's `resumeSchema`
+   - Automatically resumes the primitive with the extracted data
+
+### Example
+
+```typescript
+const stream = await routingAgent.network("Delete the old records", {
+  autoResumeSuspendedTools: true,
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+
+for await (const chunk of stream) {
+  if (chunk.type === "workflow-execution-suspended") {
+    console.log(chunk.payload.suspendPayload);
+    // { message: "Please confirm: delete old records", action: "delete old records" }
+  }
+}
+
+// User provides confirmation in their next message
+const resumedStream = await routingAgent.network("Yes, confirmed", {
+  autoResumeSuspendedTools: true,
+  memory: {
+    thread: "user-123",
+    resource: "my-app"
+  }
+});
+
+for await (const chunk of resumedStream) {
+  if (chunk.type === "network-execution-event-step-finish") {
+    console.log(chunk.payload.result);
+  }
+}
+```
+
+**Conversation flow:**
+
+```
+User: "Delete the old records"
+Agent: "Please confirm: delete old records"
+
+User: "Yes, confirmed"
+Agent: "Records deleted successfully"
+```
+
+### Requirements
+
+For automatic tool resumption to work:
+
+- **Memory configured**: The agent needs memory to track suspended tools across messages
+- **Same thread**: The follow-up message must use the same memory thread and resource identifiers
+- **`resumeSchema` defined**: The tool (either directly in the network agent or in a subAgent) / workflow (step that gets suspended) must define a `resumeSchema` so the agent knows what data to extract from the user's message
+
+### Manual vs automatic resumption
+
+| Approach | Use case |
+|----------|----------|
+| Manual (`resumeNetwork()`) | Programmatic control, webhooks, button clicks, external triggers |
+| Automatic (`autoResumeSuspendedTools`) | Conversational flows where users provide resume data in natural language |
+
+Both approaches work with the same tool definitions. Automatic resumption triggers only when suspended tools exist in the message history and the user sends a new message on the same thread.
+
+## Related
+
+- [Agent Networks](./networks)
+- [Agent Approval](./agent-approval)
+- [Human-in-the-Loop](/docs/v1/workflows/human-in-the-loop)
+- [Agent Memory](./agent-memory)