fix: enhance JSON parsing to handle multiple objects and add corresponding tests

Author: yaonyan

deno.lock

@@ -1,6 +1,6 @@
 {
   "name": "@mcpc/cli",
-  "version": "0.1.6",
+  "version": "0.1.7",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/mcpc-tech/mcpc.git"

packages/cli/deno.json

@@ -1,6 +1,6 @@
 {
   "name": "@mcpc/core",
-  "version": "0.2.5",
+  "version": "0.2.6",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/mcpc-tech/mcpc.git"

packages/core/deno.json

@@ -189,10 +189,10 @@ export class SamplingExecutor extends BaseSamplingExecutor {
     const taskPrompt = `
 
 ## Current Task
-I will now use agentic sampling to complete the following task: "${userRequest}"${contextInfo}
+You will now use agentic sampling to complete the following task: "${userRequest}"${contextInfo}
 
-When I need to use a tool, I should specify the tool name in 'action' and provide tool-specific parameters as additional properties.
-When the task is complete, I should use "action": "complete".`;
+When you need to use a tool, specify the tool name in 'action' and provide tool-specific parameters as additional properties.
+When the task is complete, use "action": "complete".`;
 
     // Use JSON instruction injection pattern
     return this.injectJsonInstruction({

packages/core/src/executors/sampling/agentic-sampling-executor.ts

@@ -1,7 +1,6 @@
 import type { CallToolResult } from "@modelcontextprotocol/sdk/types.js";
 import type { ComposableMCPServer } from "../../compose.ts";
 import type { SamplingConfig } from "../../types.ts";
-import { CompiledPrompts } from "../../prompts/index.ts";
 import { Ajv } from "ajv";
 import { AggregateAjvError } from "@segment/ajv-human-errors";
 import addFormats from "ajv-formats";
@@ -103,7 +102,7 @@ export abstract class BaseSamplingExecutor {
       content: {
         type: "text",
         text:
-          'Return ONLY raw JSON (no code fences or explanations). The JSON MUST include action and decision. Example: {"action":"<tool>","decision":"proceed|complete","<tool>":{}}',
+          'Return ONE AND ONLY ONE raw JSON object (no code fences, explanations, or multiple objects). The JSON MUST include action and decision. Example: {"action":"<tool>","decision":"proceed|complete","<tool>":{}}',
       },
     }];
 
@@ -160,15 +159,14 @@ export abstract class BaseSamplingExecutor {
             continue;
           }
 
-          if (parsedData) {
-            this.conversationHistory.push({
-              role: "assistant",
-              content: {
-                type: "text",
-                text: JSON.stringify(parsedData, null, 2),
-              },
-            });
-          }
+          // Always show LLM what we parsed - this allows self-correction
+          this.conversationHistory.push({
+            role: "assistant",
+            content: {
+              type: "text",
+              text: JSON.stringify(parsedData, null, 2),
+            },
+          });
 
           const action = parsedData["action"];
 
@@ -285,28 +283,18 @@ export abstract class BaseSamplingExecutor {
   }
 
   protected addParsingErrorToHistory(
-    responseText: string,
+    _responseText: string,
     parseError: unknown,
   ): void {
-    this.conversationHistory.push({
-      role: "assistant",
-      content: {
-        type: "text",
-        text: `JSON parsing failed. Response was: ${responseText}`,
-      },
-    });
+    const errorMsg = parseError instanceof Error
+      ? parseError.message
+      : String(parseError);
 
     this.conversationHistory.push({
       role: "user",
       content: {
         type: "text",
-        text: CompiledPrompts.errorResponse({
-          errorMessage: `JSON parsing failed: ${
-            parseError instanceof Error
-              ? parseError.message
-              : String(parseError)
-          }\n\nPlease respond with valid JSON.`,
-        }),
+        text: `Invalid JSON: ${errorMsg}\n\nRespond with valid JSON.`,
       },
     });
   }
@@ -513,11 +501,11 @@ ${history}`,
     schema,
     schemaPrefix = "JSON schema:",
     schemaSuffix = `STRICT REQUIREMENTS:
-1. Return ONLY raw JSON that passes JSON.parse() - no markdown, code blocks, explanatory text, or extra characters
+1. Return ONE AND ONLY ONE raw JSON object that passes JSON.parse() - no markdown, code blocks, explanatory text, or multiple JSON objects
 2. Include ALL required fields with correct data types and satisfy ALL schema constraints (anyOf, oneOf, allOf, not, enum, pattern, min/max, conditionals)
-3. Your response must be the JSON object itself, nothing else
+3. Your response must be a single JSON object, nothing else
 
-INVALID: \`\`\`json{"key":"value"}\`\`\` or "Here is: {"key":"value"}"
+INVALID: \`\`\`json{"key":"value"}\`\`\` or "Here is: {"key":"value"}" or {"key":"value"}{"key":"value"}
 VALID: {"key":"value"}`,
   }: {
     prompt?: string;

packages/core/src/executors/sampling/base-sampling-executor.ts

@@ -1,6 +1,6 @@
 {
   "name": "@mcpc/mcp-sampling-ai-provider",
-  "version": "0.1.7",
+  "version": "0.1.8",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/mcpc-tech/mcpc.git"

packages/mcp-sampling-ai-provider/deno.json

@@ -1,6 +1,6 @@
 {
   "name": "@mcpc/utils",
-  "version": "0.2.2",
+  "version": "0.2.3",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/mcpc-tech/mcpc.git"
@@ -15,6 +15,7 @@
   },
   "imports": {
     "@modelcontextprotocol/sdk": "npm:@modelcontextprotocol/sdk@^1.8.0",
+    "@std/assert": "jsr:@std/assert@^1",
     "@std/http": "jsr:@std/http@^1.0.14",
     "cheerio": "npm:cheerio@^1.0.0",
     "jsonrepair": "npm:jsonrepair@^3.13.0"

packages/utils/deno.json

@@ -17,10 +17,54 @@ function stripMarkdownAndText(text: string): string {
     "",
   );
 
-  // Try to find JSON object/array boundaries if there's surrounding text
-  const jsonMatch = text.match(/(\{[\s\S]*\}|\[[\s\S]*\])/);
-  if (jsonMatch) {
-    text = jsonMatch[1];
+  // Try to extract the FIRST complete JSON object or array
+  // This handles cases where LLM returns multiple JSON objects
+  const firstJsonIndex = text.search(/[\{\[]/);
+  if (firstJsonIndex >= 0) {
+    text = text.slice(firstJsonIndex);
+
+    // Find the end of the first complete JSON structure
+    let depth = 0;
+    let inString = false;
+    let escapeNext = false;
+    const startChar = text[0];
+    const endChar = startChar === "{" ? "}" : "]";
+
+    for (let i = 0; i < text.length; i++) {
+      const char = text[i];
+
+      if (escapeNext) {
+        escapeNext = false;
+        continue;
+      }
+
+      if (char === "\\") {
+        escapeNext = true;
+        continue;
+      }
+
+      if (char === '"' && !inString) {
+        inString = true;
+        continue;
+      }
+
+      if (char === '"' && inString) {
+        inString = false;
+        continue;
+      }
+
+      if (inString) continue;
+
+      if (char === startChar) {
+        depth++;
+      } else if (char === endChar) {
+        depth--;
+        if (depth === 0) {
+          // Found the end of the first complete JSON structure
+          return text.slice(0, i + 1);
+        }
+      }
+    }
   }
 
   return text.trim();

packages/utils/src/json.ts

@@ -0,0 +1,78 @@
+import { assertEquals } from "@std/assert";
+import { parseJSON } from "../src/json.ts";
+
+Deno.test("parseJSON - handles multiple JSON objects by taking the first one", () => {
+  const duplicateJson =
+    '{"action":"filesystem_read_file","decision":"proceed","nextAction":"filesystem_read_file","filesystem_read_file":{"path":"codex-rs/core/src/mcp_connection_manager.rs"}}{"action":"filesystem_read_file","decision":"proceed","nextAction":"filesystem_read_file","filesystem_read_file":{"path":"codex-rs/core/src/mcp_connection_manager.rs"}}';
+
+  const result = parseJSON(duplicateJson, true);
+
+  assertEquals(result, {
+    action: "filesystem_read_file",
+    decision: "proceed",
+    nextAction: "filesystem_read_file",
+    filesystem_read_file: {
+      path: "codex-rs/core/src/mcp_connection_manager.rs",
+    },
+  });
+});
+
+Deno.test("parseJSON - handles nested objects correctly", () => {
+  const nestedJson =
+    '{"outer":{"inner":{"deep":"value"},"array":[1,2,3]},"next":"field"}{"should":"ignore"}';
+
+  const result = parseJSON(nestedJson, true);
+
+  assertEquals(result, {
+    outer: {
+      inner: {
+        deep: "value",
+      },
+      array: [1, 2, 3],
+    },
+    next: "field",
+  });
+});
+
+Deno.test("parseJSON - handles JSON arrays", () => {
+  const arrayJson = '[{"a":1},{"b":2}][{"c":3}]';
+
+  const result = parseJSON(arrayJson, true);
+
+  assertEquals(result, [{ a: 1 }, { b: 2 }]);
+});
+
+Deno.test("parseJSON - handles strings with quotes correctly", () => {
+  const stringWithQuotes =
+    '{"message":"He said \\"hello\\" to me"}{"ignore":"this"}';
+
+  const result = parseJSON(stringWithQuotes, true);
+
+  assertEquals(result, {
+    message: 'He said "hello" to me',
+  });
+});
+
+Deno.test("parseJSON - handles markdown code fences with multiple objects", () => {
+  const markdownJson =
+    '```json\n{"action":"test","value":123}{"extra":"object"}\n```';
+
+  const result = parseJSON(markdownJson, true);
+
+  assertEquals(result, {
+    action: "test",
+    value: 123,
+  });
+});
+
+Deno.test("parseJSON - handles explanatory text before JSON", () => {
+  const textBeforeJson =
+    'Here is the response: {"action":"test","value":123}{"extra":"object"}';
+
+  const result = parseJSON(textBeforeJson, true);
+
+  assertEquals(result, {
+    action: "test",
+    value: 123,
+  });
+});