@probelabs/probe 0.6.0-rc238 → 0.6.0-rc240

package/build/tools/executePlan.js

@@ -6,7 +6,7 @@
  */
 
 import { tool } from 'ai';
-import { executePlanSchema, parseAndResolvePaths } from './common.js';
+import { executePlanSchema, cleanupExecutePlanSchema, parseAndResolvePaths } from './common.js';
 import { createDSLRuntime } from '../agent/dsl/runtime.js';
 import { search } from '../search.js';
 import { query } from '../query.js';
@@ -347,7 +347,16 @@ RULES REMINDER:
           });
         }
 
-        // All retries exhausted
+        // All retries exhausted — auto-cleanup output buffer to prevent stale data
+        if (outputBuffer && outputBuffer.items && outputBuffer.items.length > 0) {
+          const clearedChars = outputBuffer.items.reduce((sum, item) => sum + item.length, 0);
+          outputBuffer.items = [];
+          planSpan?.addEvent?.('dsl.auto_cleanup', {
+            'cleanup.chars_cleared': clearedChars,
+            'cleanup.reason': 'all_retries_exhausted',
+          });
+        }
+
         finalOutput = `Plan execution failed after ${maxRetries} retries.\n\nLast error: ${lastError}`;
         planSpan?.setAttributes?.({
           'dsl.result': 'all_retries_exhausted',
@@ -358,6 +367,11 @@ RULES REMINDER:
         planSpan?.end?.();
         return finalOutput;
       } catch (e) {
+        // Auto-cleanup output buffer on exception to prevent stale data
+        if (outputBuffer && outputBuffer.items && outputBuffer.items.length > 0) {
+          outputBuffer.items = [];
+        }
+
         planSpan?.setStatus?.('ERROR');
         planSpan?.addEvent?.('exception', {
           'exception.message': e.message,
@@ -808,3 +822,87 @@ output(table);
 return "Generated table with " + results.length + " items.";
 \`\`\``;
 }
+
+/**
+ * Create the cleanup_execute_plan tool for the Vercel AI SDK.
+ *
+ * Cleans up output buffer and optionally session store from previous
+ * failed or interrupted execute_plan calls.
+ *
+ * @param {Object} options
+ * @param {Object} [options.outputBuffer] - Output buffer to clear
+ * @param {Object} [options.sessionStore] - Session store to clear
+ * @param {Object} [options.tracer] - OTEL tracer for tracing
+ * @returns {Object} Vercel AI SDK tool
+ */
+export function createCleanupExecutePlanTool(options) {
+  const { outputBuffer, sessionStore, tracer } = options;
+
+  return tool({
+    description: 'Clean up output buffer and session store from previous execute_plan calls. ' +
+      'Use this when a previous execute_plan failed and left stale data, or before starting a fresh analysis.',
+    parameters: cleanupExecutePlanSchema,
+    execute: async ({ clearOutputBuffer = true, clearSessionStore = false }) => {
+      const span = tracer?.createToolSpan?.('cleanup_execute_plan', {
+        'cleanup.clear_output_buffer': clearOutputBuffer,
+        'cleanup.clear_session_store': clearSessionStore,
+      }) || null;
+
+      const results = [];
+
+      try {
+        if (clearOutputBuffer && outputBuffer) {
+          const itemCount = outputBuffer.items?.length || 0;
+          const charCount = outputBuffer.items?.reduce((sum, item) => sum + item.length, 0) || 0;
+          outputBuffer.items = [];
+          results.push(`Output buffer cleared (${itemCount} items, ${charCount} chars)`);
+        }
+
+        if (clearSessionStore && sessionStore) {
+          const keyCount = Object.keys(sessionStore).length;
+          for (const key of Object.keys(sessionStore)) {
+            delete sessionStore[key];
+          }
+          results.push(`Session store cleared (${keyCount} keys)`);
+        }
+
+        const output = results.length > 0
+          ? `Cleanup complete:\n- ${results.join('\n- ')}`
+          : 'Nothing to clean up';
+
+        span?.setAttributes?.({
+          'cleanup.result': output,
+          'cleanup.success': true,
+        });
+        span?.setStatus?.('OK');
+        span?.end?.();
+
+        return output;
+      } catch (e) {
+        span?.setStatus?.('ERROR');
+        span?.addEvent?.('exception', { 'exception.message': e.message });
+        span?.end?.();
+        return `Cleanup failed: ${e.message}`;
+      }
+    },
+  });
+}
+
+/**
+ * XML tool definition for cleanup_execute_plan.
+ *
+ * @returns {string} Tool definition text
+ */
+export function getCleanupExecutePlanToolDefinition() {
+  return `## cleanup_execute_plan
+Description: Clean up output buffer and session store from previous execute_plan calls. Use when a previous execute_plan failed and left stale data, or before starting a fresh analysis.
+
+Parameters:
+- clearOutputBuffer: (optional, default: true) Clear the output buffer from previous execute_plan calls
+- clearSessionStore: (optional, default: false) Clear the session store (persisted data across execute_plan calls)
+
+Example:
+<cleanup_execute_plan>
+<clearOutputBuffer>true</clearOutputBuffer>
+</cleanup_execute_plan>`;
+}

package/build/tools/index.js

@@ -11,8 +11,8 @@ export { editTool, createTool } from './edit.js';
 // Export LangChain tools
 export { createSearchTool, createQueryTool, createExtractTool } from './langchain.js';
 
-// Export execute_plan tool
-export { createExecutePlanTool, getExecutePlanToolDefinition } from './executePlan.js';
+// Export execute_plan and cleanup_execute_plan tools
+export { createExecutePlanTool, getExecutePlanToolDefinition, createCleanupExecutePlanTool, getCleanupExecutePlanToolDefinition } from './executePlan.js';
 
 // Export common schemas and utilities
 export {
@@ -22,6 +22,7 @@ export {
 	delegateSchema,
 	bashSchema,
 	executePlanSchema,
+	cleanupExecutePlanSchema,
 	delegateDescription,
 	delegateToolDefinition,
 	bashDescription,

package/cjs/agent/ProbeAgent.cjs

@@ -39352,7 +39352,7 @@ function resolveTargetPath(target, cwd) {
   }
   return filePart + suffix;
 }
-var import_path6, searchSchema, querySchema, extractSchema, delegateSchema, listSkillsSchema, useSkillSchema, bashSchema, analyzeAllSchema, executePlanSchema, attemptCompletionSchema, searchToolDefinition, queryToolDefinition, extractToolDefinition, delegateToolDefinition, attemptCompletionToolDefinition, analyzeAllToolDefinition, bashToolDefinition, googleSearchToolDefinition, urlContextToolDefinition, searchDescription, queryDescription, extractDescription, delegateDescription, analyzeAllDescription, DEFAULT_VALID_TOOLS;
+var import_path6, searchSchema, querySchema, extractSchema, delegateSchema, listSkillsSchema, useSkillSchema, bashSchema, analyzeAllSchema, executePlanSchema, cleanupExecutePlanSchema, attemptCompletionSchema, searchToolDefinition, queryToolDefinition, extractToolDefinition, delegateToolDefinition, attemptCompletionToolDefinition, analyzeAllToolDefinition, bashToolDefinition, googleSearchToolDefinition, urlContextToolDefinition, searchDescription, queryDescription, extractDescription, delegateDescription, analyzeAllDescription, DEFAULT_VALID_TOOLS;
 var init_common2 = __esm({
   "src/tools/common.js"() {
     "use strict";
@@ -39401,6 +39401,10 @@ var init_common2 = __esm({
       code: external_exports.string().min(1).describe("JavaScript DSL code to execute. All function calls look synchronous \u2014 do NOT use async/await. Use map(items, fn) for batch operations. Use LLM(instruction, data) for AI processing."),
       description: external_exports.string().optional().describe("Human-readable description of what this plan does, for logging.")
     });
+    cleanupExecutePlanSchema = external_exports.object({
+      clearOutputBuffer: external_exports.boolean().optional().default(true).describe("Clear the output buffer from previous execute_plan calls"),
+      clearSessionStore: external_exports.boolean().optional().default(false).describe("Clear the session store (persisted data across execute_plan calls)")
+    });
     attemptCompletionSchema = {
       // Custom validation that requires result parameter but allows direct XML response
       safeParse: (params) => {
@@ -39748,6 +39752,7 @@ Capabilities:
       "delegate",
       "analyze_all",
       "execute_plan",
+      "cleanup_execute_plan",
       "listSkills",
       "useSkill",
       "listFiles",
@@ -50948,9 +50953,6 @@ function validateDSL(code) {
     if (node.type === "FunctionExpression" && node.generator) {
       errors.push(`Generator functions are not allowed at position ${node.start}`);
     }
-    if (node.type === "Literal" && node.regex) {
-      errors.push(`Regex literals are not supported at position ${node.start}. Use String methods like indexOf(), includes(), startsWith() instead.`);
-    }
     if (node.type === "Identifier" && BLOCKED_IDENTIFIERS.has(node.name)) {
       errors.push(`Blocked identifier: '${node.name}' at position ${node.start}`);
     }
@@ -50989,18 +50991,23 @@ var init_validator = __esm({
       "BlockStatement",
       "VariableDeclaration",
       "VariableDeclarator",
+      "FunctionDeclaration",
       "ArrowFunctionExpression",
       "FunctionExpression",
       "CallExpression",
+      "NewExpression",
       "MemberExpression",
       "Identifier",
       "Literal",
       "TemplateLiteral",
       "TemplateElement",
+      "TaggedTemplateExpression",
       "ArrayExpression",
       "ObjectExpression",
       "SpreadElement",
       "IfStatement",
+      "SwitchStatement",
+      "SwitchCase",
       "ConditionalExpression",
       "ForOfStatement",
       "ForInStatement",
@@ -51462,7 +51469,6 @@ function createDSLRuntime(options) {
     mcpTools = {},
     llmCall,
     mapConcurrency = 3,
-    timeoutMs = 12e4,
     maxLoopIterations = 5e3,
     tracer = null,
     sessionStore = {},
@@ -51517,7 +51523,6 @@ ${validation.errors.join("\n")}`,
       };
     }
     tracer?.addEvent?.("dsl.phase.execute_start", {
-      "dsl.timeout_ms": timeoutMs,
       "dsl.max_loop_iterations": maxLoopIterations
     });
     try {
@@ -51545,18 +51550,10 @@ ${validation.errors.join("\n")}`,
         escapedError = reason;
       };
       process.on("unhandledRejection", rejectionHandler);
-      let timeoutHandle;
-      const executionPromise = exec6().run();
-      const timeoutPromise = new Promise((_, reject2) => {
-        timeoutHandle = setTimeout(() => {
-          reject2(new Error(`Execution timed out after ${Math.round(timeoutMs / 1e3)}s. Script took too long \u2014 reduce the amount of work (fewer items, smaller data) or increase timeout.`));
-        }, timeoutMs);
-      });
       let result;
       try {
-        result = await Promise.race([executionPromise, timeoutPromise]);
+        result = await exec6().run();
       } finally {
-        clearTimeout(timeoutHandle);
         setTimeout(() => {
           process.removeListener("unhandledRejection", rejectionHandler);
         }, 500);
@@ -58542,6 +58539,14 @@ Logs: ${result.logs.join(" | ")}` : "";
             "dsl.error": lastError.substring(0, 1e3)
           });
         }
+        if (outputBuffer && outputBuffer.items && outputBuffer.items.length > 0) {
+          const clearedChars = outputBuffer.items.reduce((sum, item) => sum + item.length, 0);
+          outputBuffer.items = [];
+          planSpan?.addEvent?.("dsl.auto_cleanup", {
+            "cleanup.chars_cleared": clearedChars,
+            "cleanup.reason": "all_retries_exhausted"
+          });
+        }
         finalOutput = `Plan execution failed after ${maxRetries} retries.
 
 Last error: ${lastError}`;
@@ -58554,6 +58559,9 @@ Last error: ${lastError}`;
         planSpan?.end?.();
         return finalOutput;
       } catch (e5) {
+        if (outputBuffer && outputBuffer.items && outputBuffer.items.length > 0) {
+          outputBuffer.items = [];
+        }
         planSpan?.setStatus?.("ERROR");
         planSpan?.addEvent?.("exception", {
           "exception.message": e5.message,
@@ -58997,6 +59005,62 @@ output(table);
 return "Generated table with " + results.length + " items.";
 \`\`\``;
 }
+function createCleanupExecutePlanTool(options) {
+  const { outputBuffer, sessionStore, tracer } = options;
+  return (0, import_ai4.tool)({
+    description: "Clean up output buffer and session store from previous execute_plan calls. Use this when a previous execute_plan failed and left stale data, or before starting a fresh analysis.",
+    parameters: cleanupExecutePlanSchema,
+    execute: async ({ clearOutputBuffer = true, clearSessionStore = false }) => {
+      const span = tracer?.createToolSpan?.("cleanup_execute_plan", {
+        "cleanup.clear_output_buffer": clearOutputBuffer,
+        "cleanup.clear_session_store": clearSessionStore
+      }) || null;
+      const results = [];
+      try {
+        if (clearOutputBuffer && outputBuffer) {
+          const itemCount = outputBuffer.items?.length || 0;
+          const charCount = outputBuffer.items?.reduce((sum, item) => sum + item.length, 0) || 0;
+          outputBuffer.items = [];
+          results.push(`Output buffer cleared (${itemCount} items, ${charCount} chars)`);
+        }
+        if (clearSessionStore && sessionStore) {
+          const keyCount = Object.keys(sessionStore).length;
+          for (const key of Object.keys(sessionStore)) {
+            delete sessionStore[key];
+          }
+          results.push(`Session store cleared (${keyCount} keys)`);
+        }
+        const output = results.length > 0 ? `Cleanup complete:
+- ${results.join("\n- ")}` : "Nothing to clean up";
+        span?.setAttributes?.({
+          "cleanup.result": output,
+          "cleanup.success": true
+        });
+        span?.setStatus?.("OK");
+        span?.end?.();
+        return output;
+      } catch (e5) {
+        span?.setStatus?.("ERROR");
+        span?.addEvent?.("exception", { "exception.message": e5.message });
+        span?.end?.();
+        return `Cleanup failed: ${e5.message}`;
+      }
+    }
+  });
+}
+function getCleanupExecutePlanToolDefinition() {
+  return `## cleanup_execute_plan
+Description: Clean up output buffer and session store from previous execute_plan calls. Use when a previous execute_plan failed and left stale data, or before starting a fresh analysis.
+
+Parameters:
+- clearOutputBuffer: (optional, default: true) Clear the output buffer from previous execute_plan calls
+- clearSessionStore: (optional, default: false) Clear the session store (persisted data across execute_plan calls)
+
+Example:
+<cleanup_execute_plan>
+<clearOutputBuffer>true</clearOutputBuffer>
+</cleanup_execute_plan>`;
+}
 var import_ai4;
 var init_executePlan = __esm({
   "src/tools/executePlan.js"() {
@@ -59352,6 +59416,13 @@ function createWrappedTools(baseTools) {
       baseTools.executePlanTool.execute
     );
   }
+  if (baseTools.cleanupExecutePlanTool) {
+    wrappedTools.cleanupExecutePlanToolInstance = wrapToolWithEmitter(
+      baseTools.cleanupExecutePlanTool,
+      "cleanup_execute_plan",
+      baseTools.cleanupExecutePlanTool.execute
+    );
+  }
   if (baseTools.bashTool) {
     wrappedTools.bashToolInstance = wrapToolWithEmitter(
       baseTools.bashTool,
@@ -60268,6 +60339,9 @@ function createTools(configOptions) {
   }
   if (configOptions.enableExecutePlan && isToolAllowed("execute_plan")) {
     tools2.executePlanTool = createExecutePlanTool(configOptions);
+    if (isToolAllowed("cleanup_execute_plan")) {
+      tools2.cleanupExecutePlanTool = createCleanupExecutePlanTool(configOptions);
+    }
   } else if (isToolAllowed("analyze_all")) {
     tools2.analyzeAllTool = analyzeAllTool(configOptions);
   }
@@ -97681,15 +97755,31 @@ function isSimpleTextWrapperSchema(schema) {
     return null;
   }
   const trimmed = schema.trim();
-  const simplePatterns = [
-    /^\{\s*["']?(\w+)["']?\s*:\s*["']?string["']?\s*\}$/i,
-    /^\{\s*["']?type["']?\s*:\s*["']?object["']?\s*,\s*["']?properties["']?\s*:\s*\{\s*["']?(\w+)["']?\s*:\s*\{\s*["']?type["']?\s*:\s*["']?string["']?\s*\}\s*\}\s*\}$/i
-  ];
-  for (const pattern of simplePatterns) {
-    const match2 = trimmed.match(pattern);
-    if (match2) {
-      return { fieldName: match2[1] };
+  try {
+    const parsed = JSON.parse(trimmed);
+    if (typeof parsed !== "object" || parsed === null) {
+    } else {
+      const keys2 = Object.keys(parsed);
+      if (keys2.length === 1 && parsed[keys2[0]] === "string") {
+        return { fieldName: keys2[0] };
+      }
+      if (parsed.type === "object" && parsed.properties) {
+        const propKeys = Object.keys(parsed.properties);
+        if (propKeys.length === 1) {
+          const prop = parsed.properties[propKeys[0]];
+          if (prop && prop.type === "string") {
+            return { fieldName: propKeys[0] };
+          }
+        }
+      }
+      return null;
     }
+  } catch {
+  }
+  const simplePattern = /^\{\s*["']?(\w+)["']?\s*:\s*["']?string["']?\s*\}$/i;
+  const match2 = trimmed.match(simplePattern);
+  if (match2) {
+    return { fieldName: match2[1] };
   }
   return null;
 }
@@ -110551,6 +110641,9 @@ var init_ProbeAgent = __esm({
         }
         if (this.enableExecutePlan && wrappedTools.executePlanToolInstance && isToolAllowed("execute_plan")) {
           this.toolImplementations.execute_plan = wrappedTools.executePlanToolInstance;
+          if (wrappedTools.cleanupExecutePlanToolInstance && isToolAllowed("cleanup_execute_plan")) {
+            this.toolImplementations.cleanup_execute_plan = wrappedTools.cleanupExecutePlanToolInstance;
+          }
         } else if (wrappedTools.analyzeAllToolInstance && isToolAllowed("analyze_all")) {
           this.toolImplementations.analyze_all = wrappedTools.analyzeAllToolInstance;
         }
@@ -111924,6 +112017,10 @@ Workspace: ${this.allowedFolders.join(", ")}`;
           if (this.enableBash && isToolAllowed("bash")) dslFunctions.push("bash");
           toolDefinitions += `${getExecutePlanToolDefinition(dslFunctions)}
 `;
+          if (isToolAllowed("cleanup_execute_plan")) {
+            toolDefinitions += `${getCleanupExecutePlanToolDefinition()}
+`;
+          }
         } else if (isToolAllowed("analyze_all")) {
           toolDefinitions += `${analyzeAllToolDefinition}
 `;
@@ -111999,6 +112096,9 @@ The configuration is loaded from src/config.js lines 15-25 which contains the da
         }
         if (this.enableExecutePlan && isToolAllowed("execute_plan")) {
           availableToolsList += '- execute_plan: Execute a DSL program to orchestrate tool calls. ALWAYS use this for: questions containing "all"/"every"/"comprehensive"/"complete inventory", multi-topic analysis, open-ended discovery questions, or any task requiring full codebase coverage.\n';
+          if (isToolAllowed("cleanup_execute_plan")) {
+            availableToolsList += "- cleanup_execute_plan: Clean up output buffer and session store from previous execute_plan calls.\n";
+          }
         } else if (isToolAllowed("analyze_all")) {
           availableToolsList += "- analyze_all: Process ALL data matching a query using map-reduce (for aggregate questions needing 100% coverage).\n";
         }
@@ -112225,7 +112325,7 @@ You are working with a workspace. Available paths: ${workspaceDesc}
         }
         try {
           const oldHistoryLength = this.history.length;
-          if (this._outputBuffer) {
+          if (this._outputBuffer && !options?._schemaFormatted) {
             this._outputBuffer.items = [];
           }
           if (this.enableTasks) {
@@ -112585,6 +112685,9 @@ You are working with a workspace. Available paths: ${workspaceDesc}
               }
               if (this.enableExecutePlan && this.allowedTools.isEnabled("execute_plan")) {
                 validTools.push("execute_plan");
+                if (this.allowedTools.isEnabled("cleanup_execute_plan")) {
+                  validTools.push("cleanup_execute_plan");
+                }
               } else if (this.allowedTools.isEnabled("analyze_all")) {
                 validTools.push("analyze_all");
               }