wave-agent-sdk 0.5.0 → 0.6.0

package/src/services/taskManager.ts

@@ -0,0 +1,188 @@
+import { promises as fs } from "fs";
+import { join } from "path";
+import { homedir } from "os";
+import { EventEmitter } from "events";
+import { Task } from "../types/tasks.js";
+import { logger } from "../utils/globalLogger.js";
+
+export class TaskManager extends EventEmitter {
+  private readonly baseDir: string;
+  private taskListId: string;
+
+  constructor(taskListId: string) {
+    super();
+    this.taskListId = taskListId;
+    this.baseDir = join(homedir(), ".wave", "tasks");
+  }
+
+  public getTaskListId(): string {
+    return this.taskListId;
+  }
+
+  public setTaskListId(taskListId: string): void {
+    this.taskListId = taskListId;
+  }
+
+  private getSessionDir(): string {
+    return join(this.baseDir, this.taskListId);
+  }
+
+  private getTaskPath(taskId: string): string {
+    return join(this.getSessionDir(), `${taskId}.json`);
+  }
+
+  private getLockPath(): string {
+    return join(this.getSessionDir(), `.lock`);
+  }
+
+  async ensureSessionDir(): Promise<void> {
+    await fs.mkdir(this.getSessionDir(), { recursive: true });
+  }
+
+  private async withLock<T>(operation: () => Promise<T>): Promise<T> {
+    const lockPath = this.getLockPath();
+    let lockHandle;
+    const maxRetries = 100;
+    const retryDelay = process.env.NODE_ENV === "test" ? 10 : 100;
+
+    await this.ensureSessionDir();
+
+    for (let i = 0; i < maxRetries; i++) {
+      try {
+        lockHandle = await fs.open(lockPath, "wx");
+        break;
+      } catch (error) {
+        if ((error as NodeJS.ErrnoException).code === "EEXIST") {
+          if (i === maxRetries - 1) {
+            throw new Error(
+              `Could not acquire lock for task list ${this.taskListId} after ${maxRetries} retries`,
+            );
+          }
+          await new Promise((resolve) => setTimeout(resolve, retryDelay));
+          continue;
+        }
+        throw error;
+      }
+    }
+
+    try {
+      return await operation();
+    } finally {
+      if (lockHandle) {
+        await lockHandle.close();
+      }
+      try {
+        await fs.unlink(lockPath);
+      } catch (error) {
+        logger.error(
+          `Failed to release lock for task list ${this.taskListId}:`,
+          error,
+        );
+      }
+    }
+  }
+
+  private validateTask(task: Task): void {
+    if (!task.id || typeof task.id !== "string")
+      throw new Error("Invalid task ID");
+    if (!task.subject || typeof task.subject !== "string")
+      throw new Error("Invalid task subject");
+    if (!task.description || typeof task.description !== "string")
+      throw new Error("Invalid task description");
+    if (
+      !task.status ||
+      !["pending", "in_progress", "completed", "deleted"].includes(task.status)
+    ) {
+      throw new Error(`Invalid task status: ${task.status}`);
+    }
+  }
+
+  async createTask(task: Omit<Task, "id">): Promise<string> {
+    return await this.withLock(async () => {
+      const taskId = await this.getNextTaskId();
+      const fullTask: Task = { ...task, id: taskId };
+      this.validateTask(fullTask);
+      const taskPath = this.getTaskPath(taskId);
+      const content = JSON.stringify(fullTask, null, 2);
+      await fs.writeFile(taskPath, content, "utf8");
+      this.emit("tasksChange", this.taskListId);
+      logger.info(`Task ${taskId} created in task list ${this.taskListId}`);
+      return taskId;
+    });
+  }
+
+  async getTask(taskId: string): Promise<Task | null> {
+    const taskPath = this.getTaskPath(taskId);
+    try {
+      const content = await fs.readFile(taskPath, "utf8");
+      try {
+        return JSON.parse(content.trim()) as Task;
+      } catch (parseError) {
+        logger.error(`Failed to parse task file ${taskPath}:`, parseError);
+        logger.error(`Corrupted content: ${content}`);
+        throw parseError;
+      }
+    } catch (error) {
+      if ((error as NodeJS.ErrnoException).code === "ENOENT") {
+        return null;
+      }
+      throw error;
+    }
+  }
+
+  async updateTask(task: Task): Promise<void> {
+    this.validateTask(task);
+    await this.withLock(async () => {
+      const taskPath = this.getTaskPath(task.id);
+      const content = JSON.stringify(task, null, 2);
+      await fs.writeFile(taskPath, content, "utf8");
+      this.emit("tasksChange", this.taskListId);
+      logger.info(`Task ${task.id} updated in task list ${this.taskListId}`);
+    });
+  }
+
+  async listTasks(): Promise<Task[]> {
+    const sessionDir = this.getSessionDir();
+    try {
+      const files = await fs.readdir(sessionDir);
+      const taskFiles = files.filter((f) => f.endsWith(".json"));
+
+      const tasks = await Promise.all(
+        taskFiles.map(async (file) => {
+          const taskPath = join(sessionDir, file);
+          try {
+            const content = await fs.readFile(taskPath, "utf8");
+            return JSON.parse(content.trim()) as Task;
+          } catch (error) {
+            logger.error(
+              `Failed to read or parse task file ${taskPath}:`,
+              error,
+            );
+            return null;
+          }
+        }),
+      );
+
+      return tasks.filter((t): t is Task => t !== null);
+    } catch (error) {
+      if ((error as NodeJS.ErrnoException).code === "ENOENT") {
+        return [];
+      }
+      throw error;
+    }
+  }
+
+  async getNextTaskId(): Promise<string> {
+    const tasks = await this.listTasks();
+    if (tasks.length === 0) {
+      return "1";
+    }
+    const ids = tasks.map((t) => parseInt(t.id, 10)).filter((id) => !isNaN(id));
+
+    if (ids.length === 0) {
+      return (tasks.length + 1).toString();
+    }
+
+    return (Math.max(...ids) + 1).toString();
+  }
+}

package/src/tools/askUserQuestion.ts

@@ -1,9 +1,6 @@
 import { ToolPlugin } from "./types.js";
 import { AskUserQuestionInput } from "../types/tools.js";
-import {
-  ASK_USER_QUESTION_TOOL_NAME,
-  EXIT_PLAN_MODE_TOOL_NAME,
-} from "../constants/tools.js";
+import { ASK_USER_QUESTION_TOOL_NAME } from "../constants/tools.js";
 
 export const askUserQuestionTool: ToolPlugin = {
   name: ASK_USER_QUESTION_TOOL_NAME,
@@ -11,19 +8,8 @@ export const askUserQuestionTool: ToolPlugin = {
     type: "function",
     function: {
       name: ASK_USER_QUESTION_TOOL_NAME,
-      description: `Asks the user multiple choice questions to gather information, clarify ambiguity, understand preferences, make decisions or offer them choices.
-Use this tool when you need to ask the user questions during execution. This allows you to:
-1. Gather user preferences or requirements
-2. Clarify ambiguous instructions
-3. Get decisions on implementation choices as you work
-4. Offer choices to the user about what direction to take.
-
-Usage notes:
-- Users will always be able to select "Other" to provide custom text input
-- Use multiSelect: true to allow multiple answers to be selected for a question
-- If you recommend a specific option, make that the first option in the list and add "(Recommended)" at the end of the label
-
-Plan mode note: In plan mode, use this tool to clarify requirements or choose between approaches BEFORE finalizing your plan. Do NOT use this tool to ask "Is my plan ready?" or "Should I proceed?" - use ${EXIT_PLAN_MODE_TOOL_NAME} for plan approval.`,
+      description:
+        "Asks the user multiple choice questions to gather information, clarify ambiguity, understand preferences, make decisions or offer them choices.",
       parameters: {
         type: "object",
         properties: {
@@ -31,37 +17,38 @@ Plan mode note: In plan mode, use this tool to clarify requirements or choose be
             type: "array",
             minItems: 1,
             maxItems: 4,
+            description: "Questions to ask the user (1-4 questions)",
             items: {
               type: "object",
               properties: {
                 question: {
                   type: "string",
-                  description: "The complete question to ask the user.",
+                  description:
+                    'The complete question to ask the user. Should be clear, specific, and end with a question mark. Example: "Which library should we use for date formatting?" If multiSelect is true, phrase it accordingly, e.g. "Which features do you want to enable?"',
                 },
                 header: {
                   type: "string",
                   maxLength: 12,
-                  description:
-                    "Very short label displayed as a chip/tag (max 12 chars).",
+                  description: `Very short label displayed as a chip/tag (max 12 chars). Examples: "Auth method", "Library", "Approach".`,
                 },
                 options: {
                   type: "array",
                   minItems: 2,
                   maxItems: 4,
+                  description:
+                    "The available choices for this question. Must have 2-4 options. Each option should be a distinct, mutually exclusive choice (unless multiSelect is enabled). There should be no 'Other' option, that will be provided automatically.",
                   items: {
                     type: "object",
                     properties: {
                       label: {
                         type: "string",
-                        description: "The display text for this option.",
+                        description:
+                          "The display text for this option that the user will see and select. Should be concise (1-5 words) and clearly describe the choice.",
                       },
                       description: {
                         type: "string",
-                        description: "Explanation of what this option means.",
-                      },
-                      isRecommended: {
-                        type: "boolean",
-                        description: "Whether this option is recommended.",
+                        description:
+                          "Explanation of what this option means or what will happen if chosen. Useful for providing context about trade-offs or implications.",
                       },
                     },
                     required: ["label"],
@@ -70,19 +57,54 @@ Plan mode note: In plan mode, use this tool to clarify requirements or choose be
                 multiSelect: {
                   type: "boolean",
                   default: false,
-                  description: "Allow multiple answers to be selected.",
+                  description:
+                    "Set to true to allow the user to select multiple options instead of just one. Use when choices are not mutually exclusive.",
                 },
               },
               required: ["question", "header", "options"],
             },
           },
+          answers: {
+            type: "object",
+            additionalProperties: { type: "string" },
+            description: "User answers collected by the permission component",
+          },
+          metadata: {
+            type: "object",
+            properties: {
+              source: {
+                type: "string",
+                description:
+                  'Optional identifier for the source of this question (e.g., "remember" for /remember command). Used for analytics tracking.',
+              },
+            },
+            description: "Optional metadata for the question",
+          },
         },
         required: ["questions"],
       },
     },
   },
+  prompt:
+    () => `Use this tool when you need to ask the user questions during execution. This allows you to:
+1. Gather user preferences or requirements
+2. Clarify ambiguous instructions
+3. Get decisions on implementation choices as you work
+4. Offer choices to the user about what direction to take.
+
+Usage notes:
+- Users will always be able to select "Other" to provide custom text input
+- Use multiSelect: true to allow multiple answers to be selected for a question
+- If you recommend a specific option, make that the first option in the list and add "(Recommended)" at the end of the label
+
+Plan mode note: In plan mode, use this tool to clarify requirements or choose between approaches BEFORE finalizing your plan. Do NOT use this tool to ask "Is my plan ready?" or "Should I proceed?" - use ExitPlanMode for plan approval.
+`,
   execute: async (args, context) => {
-    const { questions } = args as unknown as AskUserQuestionInput;
+    const {
+      questions,
+      answers: existingAnswers,
+      metadata,
+    } = args as unknown as AskUserQuestionInput;
 
     if (!context.permissionManager) {
       throw new Error(
@@ -94,7 +116,7 @@ Plan mode note: In plan mode, use this tool to clarify requirements or choose be
       ASK_USER_QUESTION_TOOL_NAME,
       context.permissionMode || "default",
       context.canUseToolCallback,
-      { questions },
+      { questions, answers: existingAnswers, metadata },
     );
     permissionContext.hidePersistentOption = true; // Always hide persistent option for questions
 

package/src/tools/bashTool.ts

@@ -95,6 +95,9 @@ Usage notes:
       },
     },
   },
+  prompt: () => `
+- Reserve bash tools exclusively for actual system commands and terminal operations that require shell execution. NEVER use bash echo or other command-line tools to communicate thoughts, explanations, or instructions to the user. Output all communication directly in your response text instead.
+- When making multiple bash tool calls, you MUST send a single message with multiple tools calls to run the calls in parallel. For example, if you need to run "git status" and "git diff", send a single message with two tool calls in parallel.`,
   execute: async (
     args: Record<string, unknown>,
     context: ToolContext,
@@ -218,8 +221,9 @@ Usage notes:
               );
               resolve({
                 success: true,
-                content: `Command moved to background with ID: ${taskId}. Use TaskOutput tool with task_id="${taskId}" to monitor output.`,
+                content: `Command moved to background with ID: ${taskId}.`,
                 shortResult: `Process ${taskId} backgrounded`,
+                isManuallyBackgrounded: true,
               });
             } else {
               handleAbort(
@@ -291,23 +295,13 @@ Usage notes:
       // Handle abort signal from context
       if (context?.abortSignal) {
         if (context.abortSignal.aborted) {
-          if (!isBackgrounded) {
-            handleAbort();
-          }
+          handleAbort();
           return;
         }
         // Use { once: true } to prevent listener accumulation on signal reuse
-        context.abortSignal.addEventListener(
-          "abort",
-          () => {
-            if (!isBackgrounded) {
-              handleAbort();
-            }
-          },
-          {
-            once: true,
-          },
-        );
+        context.abortSignal.addEventListener("abort", () => handleAbort(), {
+          once: true,
+        });
       }
 
       child.stdout?.on("data", (data) => {

package/src/tools/editTool.ts

@@ -22,11 +22,13 @@ function formatCompactParams(
 export const editTool: ToolPlugin = {
   name: EDIT_TOOL_NAME,
   formatCompactParams,
+  prompt: () =>
+    `Performs exact string replacements in files. \n\nUsage:\n- You must use your \`${READ_TOOL_NAME}\` tool at least once in the conversation before editing. This tool will error if you attempt an edit without reading the file. \n- When editing text from read_file tool output, ensure you preserve the exact indentation (tabs/spaces) as it appears AFTER the line number prefix. The line number prefix format is: spaces + line number + tab. Everything after that tab is the actual file content to match. Never include any part of the line number prefix in the old_string or new_string.\n- ALWAYS prefer editing existing files in the codebase. NEVER write new files unless explicitly required.\n- Only use emojis if the user explicitly requests it. Avoid adding emojis to files unless asked.\n- The edit will FAIL if \`old_string\` is not unique in the file. Either provide a larger string with more surrounding context to make it unique or use \`replace_all\` to change every instance of \`old_string\`. \n- Use \`replace_all\` for replacing and renaming strings across the file. This parameter is useful if you want to rename a variable for instance.`,
   config: {
     type: "function",
     function: {
       name: EDIT_TOOL_NAME,
-      description: `Performs exact string replacements in files. \n\nUsage:\n- You must use your \`${READ_TOOL_NAME}\` tool at least once in the conversation before editing. This tool will error if you attempt an edit without reading the file. \n- When editing text from read_file tool output, ensure you preserve the exact indentation (tabs/spaces) as it appears AFTER the line number prefix. The line number prefix format is: spaces + line number + tab. Everything after that tab is the actual file content to match. Never include any part of the line number prefix in the old_string or new_string.\n- ALWAYS prefer editing existing files in the codebase. NEVER write new files unless explicitly required.\n- Only use emojis if the user explicitly requests it. Avoid adding emojis to files unless asked.\n- The edit will FAIL if \`old_string\` is not unique in the file. Either provide a larger string with more surrounding context to make it unique or use \`replace_all\` to change every instance of \`old_string\`. \n- Use \`replace_all\` for replacing and renaming strings across the file. This parameter is useful if you want to rename a variable for instance.`,
+      description: "A tool for editing files",
       parameters: {
         type: "object",
         properties: {

package/src/tools/exitPlanMode.ts

@@ -12,8 +12,7 @@ export const exitPlanModeTool: ToolPlugin = {
     type: "function",
     function: {
       name: EXIT_PLAN_MODE_TOOL_NAME,
-      description:
-        "Use this tool when you are in plan mode and have finished writing your plan to the plan file and are ready for user approval. This tool will read the plan from the file specified in the system message and present it to the user for confirmation. You should have already written your plan to that file before calling this tool.",
+      description: "Prompts the user to exit plan mode and start coding",
       parameters: {
         type: "object",
         properties: {},
@@ -22,6 +21,31 @@ export const exitPlanModeTool: ToolPlugin = {
       },
     },
   },
+  prompt:
+    () => `Use this tool when you are in plan mode and have finished writing your plan to the plan file and are ready for user approval.
+
+## How This Tool Works
+- You should have already written your plan to the plan file specified in the plan mode system message
+- This tool does NOT take the plan content as a parameter - it will read the plan from the file you wrote
+- This tool simply signals that you're done planning and ready for the user to review and approve
+- The user will see the contents of your plan file when they review it
+
+## When to Use This Tool
+IMPORTANT: Only use this tool when the task requires planning the implementation steps of a task that requires writing code. For research tasks where you're gathering information, searching files, reading files or in general trying to understand the codebase - do NOT use this tool.
+
+## Before Using This Tool
+Ensure your plan is complete and unambiguous:
+- If you have unresolved questions about requirements or approach, use AskUserQuestion first (in earlier phases)
+- Once your plan is finalized, use THIS tool to request approval
+
+**Important:** Do NOT use AskUserQuestion to ask "Is this plan okay?" or "Should I proceed?" - that's exactly what THIS tool does. ExitPlanMode inherently requests user approval of your plan.
+
+## Examples
+
+1. Initial task: "Search for and understand the implementation of vim mode in the codebase" - Do not use the exit plan mode tool because you are not planning the implementation steps of a task.
+2. Initial task: "Help me implement yank mode for vim" - Use the exit plan mode tool after you have finished planning the implementation steps of the task.
+3. Initial task: "Add a new feature to handle user authentication" - If unsure about auth method (OAuth, JWT, etc.), use AskUserQuestion first, then use exit plan mode tool after clarifying the approach.
+`,
   execute: async (
     _args: Record<string, unknown>,
     context: ToolContext,

package/src/tools/globTool.ts

@@ -3,7 +3,7 @@ import { stat } from "fs/promises";
 import type { ToolPlugin, ToolResult, ToolContext } from "./types.js";
 import { resolvePath, getDisplayPath } from "../utils/path.js";
 import { getGlobIgnorePatterns } from "../utils/fileFilter.js";
-import { GLOB_TOOL_NAME, TASK_TOOL_NAME } from "../constants/tools.js";
+import { GLOB_TOOL_NAME } from "../constants/tools.js";
 
 /**
  * Glob Tool Plugin - Fast file pattern matching
@@ -14,7 +14,8 @@ export const globTool: ToolPlugin = {
     type: "function",
     function: {
       name: GLOB_TOOL_NAME,
-      description: `- Fast file pattern matching tool that works with any codebase size\n- Supports glob patterns like "**/*.js" or "src/**/*.ts"\n- Returns matching file paths sorted by modification time\n- Use this tool when you need to find files by name patterns\n- When you are doing an open ended search that may require multiple rounds of globbing and grepping, use the ${TASK_TOOL_NAME} tool instead\n- You have the capability to call multiple tools in a single response. It is always better to speculatively perform multiple searches as a batch that are potentially useful.`,
+      description:
+        "Fast file pattern matching tool that works with any codebase size",
       parameters: {
         type: "object",
         properties: {
@@ -32,6 +33,13 @@ export const globTool: ToolPlugin = {
       },
     },
   },
+  prompt:
+    () => `- Fast file pattern matching tool that works with any codebase size
+- Supports glob patterns like "**/*.js" or "src/**/*.ts"
+- Returns matching file paths sorted by modification time
+- Use this tool when you need to find files by name patterns
+- When you are doing an open ended search that may require multiple rounds of globbing and grepping, use the Agent tool instead
+- You can call multiple tools in a single response. It is always better to speculatively perform multiple searches in parallel if they are potentially useful.`,
   execute: async (
     args: Record<string, unknown>,
     context: ToolContext,

package/src/tools/grepTool.ts

@@ -18,7 +18,7 @@ export const grepTool: ToolPlugin = {
     type: "function",
     function: {
       name: GREP_TOOL_NAME,
-      description: `A powerful search tool built on ripgrep\n\n  Usage:\n  - ALWAYS use ${GREP_TOOL_NAME} for search tasks. NEVER invoke \`grep\` or \`rg\` as a ${BASH_TOOL_NAME} command. The ${GREP_TOOL_NAME} tool has been optimized for correct permissions and access.\n  - Supports full regex syntax (e.g., "log.*Error", "function\\s+\\w+")\n  - Filter files with glob parameter (e.g., "*.js", "**/*.tsx") or type parameter (e.g., "js", "py", "rust")\n  - Output modes: "content" shows matching lines, "files_with_matches" shows only file paths (default), "count" shows match counts\n  - Use ${TASK_TOOL_NAME} tool for open-ended searches requiring multiple rounds\n  - Pattern syntax: Uses ripgrep (not grep) - literal braces need escaping (use \`interface\\{\\}\` to find \`interface{}\` in Go code)\n  - Multiline matching: By default patterns match within single lines only. For cross-line patterns like \`struct \\{[\\s\\S]*?field\`, use \`multiline: true\``,
+      description: "A powerful search tool built on ripgrep",
       parameters: {
         type: "object",
         properties: {
@@ -61,7 +61,7 @@ export const grepTool: ToolPlugin = {
           "-n": {
             type: "boolean",
             description:
-              'Show line numbers in output (rg -n). Requires output_mode: "content", ignored otherwise.',
+              'Show line numbers in output (rg -n). Requires output_mode: "content", ignored otherwise. Defaults to true.',
           },
           "-i": {
             type: "boolean",
@@ -75,7 +75,7 @@ export const grepTool: ToolPlugin = {
           head_limit: {
             type: "number",
             description:
-              'Limit output to first N lines/entries, equivalent to "| head -N". Works across all output modes: content (limits output lines), files_with_matches (limits file paths), count (limits count entries). Defaults to 100 to prevent excessive token usage.',
+              'Limit output to first N lines/entries, equivalent to "| head -N". Works across all output modes: content (limits output lines), files_with_matches (limits file paths), count (limits count entries). Defaults to 0 (unlimited).',
           },
           multiline: {
             type: "boolean",
@@ -87,6 +87,17 @@ export const grepTool: ToolPlugin = {
       },
     },
   },
+  prompt: () => `A powerful search tool built on ripgrep
+
+  Usage:
+  - ALWAYS use ${GREP_TOOL_NAME} for search tasks. NEVER invoke \`grep\` or \`rg\` as a ${BASH_TOOL_NAME} command. The ${GREP_TOOL_NAME} tool has been optimized for correct permissions and access.
+  - Supports full regex syntax (e.g., "log.*Error", "function\\s+\\w+")
+  - Filter files with glob parameter (e.g., "*.js", "**/*.tsx") or type parameter (e.g., "js", "py", "rust")
+  - Output modes: "content" shows matching lines, "files_with_matches" shows only file paths (default), "count" shows match counts
+  - Use ${TASK_TOOL_NAME} tool for open-ended searches requiring multiple rounds
+  - Pattern syntax: Uses ripgrep (not grep) - literal braces need escaping (use \`interface\\{\\}\` to find \`interface{}\` in Go code)
+  - Multiline matching: By default patterns match within single lines only. For cross-line patterns like \`struct \\{[\\s\\S]*?field\`, use \`multiline: true\`
+`,
   execute: async (
     args: Record<string, unknown>,
     context: ToolContext,
@@ -98,7 +109,7 @@ export const grepTool: ToolPlugin = {
     const contextBefore = args["-B"] as number;
     const contextAfter = args["-A"] as number;
     const contextAround = args["-C"] as number;
-    const showLineNumbers = args["-n"] as boolean;
+    const showLineNumbers = args["-n"] !== false;
     const caseInsensitive = args["-i"] as boolean;
     const fileType = args.type as string;
     const headLimit = args.head_limit as number;
@@ -218,9 +229,9 @@ export const grepTool: ToolPlugin = {
       let lines = output.split("\n");
 
       // Set default head_limit if not specified to prevent excessive token usage
-      const effectiveHeadLimit = headLimit || 100;
+      const effectiveHeadLimit = headLimit || 0;
 
-      if (lines.length > effectiveHeadLimit) {
+      if (effectiveHeadLimit > 0 && lines.length > effectiveHeadLimit) {
         lines = lines.slice(0, effectiveHeadLimit);
         finalOutput = lines.join("\n");
       }

package/src/tools/lsTool.ts

@@ -18,7 +18,7 @@ export const lsTool: ToolPlugin = {
     type: "function",
     function: {
       name: LS_TOOL_NAME,
-      description: `Lists files and directories in a given path. The path parameter must be an absolute path, not a relative path. You can optionally provide an array of glob patterns to ignore with the ignore parameter. You should generally prefer the ${GLOB_TOOL_NAME} and ${GREP_TOOL_NAME} tools, if you know which directories to search.`,
+      description: "Lists files and directories in a given path.",
       parameters: {
         type: "object",
         properties: {
@@ -39,6 +39,8 @@ export const lsTool: ToolPlugin = {
       },
     },
   },
+  prompt: () => `
+Lists files and directories in a given path. The path parameter must be an absolute path, not a relative path. You can optionally provide an array of glob patterns to ignore with the ignore parameter. You should generally prefer the ${GLOB_TOOL_NAME} and ${GREP_TOOL_NAME} tools, if you know which directories to search.`,
   execute: async (
     args: Record<string, unknown>,
     context: ToolContext,

package/src/tools/readTool.ts

@@ -131,11 +131,27 @@ async function processImageFile(
  */
 export const readTool: ToolPlugin = {
   name: READ_TOOL_NAME,
+  prompt:
+    () => `Reads a file from the local filesystem. You can access any file directly by using this tool.
+Assume this tool is able to read all files on the machine. If the User provides a path to a file assume that path is valid. It is okay to read a file that does not exist; an error will be returned.
+
+Usage:
+- The file_path parameter must be an absolute path, not a relative path
+- By default, it reads up to 2000 lines starting from the beginning of the file
+- You can optionally specify a line offset and limit (especially handy for long files), but it's recommended to read the whole file by not providing these parameters
+- Any lines longer than 2000 characters will be truncated
+- Results are returned using cat -n format, with line numbers starting at 1
+- This tool allows Agent to read images (eg PNG, JPG, etc). When reading an image file the contents are presented visually as Agent is a multimodal LLM.
+- This tool can read Jupyter notebooks (.ipynb files) and returns all cells with their outputs, combining code, text, and visualizations.
+- You have the capability to call multiple tools in a single response. It is always better to speculatively read multiple files as a batch that are potentially useful.
+- You will regularly be asked to read screenshots. If the user provides a path to a screenshot ALWAYS use this tool to view the file at the path. This tool will work with all temporary file paths like /var/folders/123/abc/T/TemporaryItems/NSIRD_screencaptureui_ZfB1tD/Screenshot.png
+- If you read a file that exists but has empty contents you will receive a system reminder warning in place of file contents.
+- Binary document formats (PDF, DOC, DOCX, XLS, XLSX, PPT, PPTX) are not supported and will return an error.`,
   config: {
     type: "function",
     function: {
       name: READ_TOOL_NAME,
-      description: `Reads a file from the local filesystem. You can access any file directly by using this tool.\nAssume this tool is able to read all files on the machine. If the User provides a path to a file assume that path is valid. It is okay to read a file that does not exist; an error will be returned.\n\nUsage:\n- The file_path parameter must be an absolute path, not a relative path\n- By default, it reads up to 2000 lines starting from the beginning of the file\n- You can optionally specify a line offset and limit (especially handy for long files), but it's recommended to read the whole file by not providing these parameters\n- Any lines longer than 2000 characters will be truncated\n- Results are returned using cat -n format, with line numbers starting at 1\n- This tool allows Agent to read images (eg PNG, JPG, etc). When reading an image file the contents are presented visually as Agent is a multimodal LLM.\n- This tool can read Jupyter notebooks (.ipynb files) and returns all cells with their outputs, combining code, text, and visualizations.\n- You have the capability to call multiple tools in a single response. It is always better to speculatively read multiple files as a batch that are potentially useful.\n- You will regularly be asked to read screenshots. If the user provides a path to a screenshot ALWAYS use this tool to view the file at the path. This tool will work with all temporary file paths like /var/folders/123/abc/T/TemporaryItems/NSIRD_screencaptureui_ZfB1tD/Screenshot.png\n- If you read a file that exists but has empty contents you will receive a system reminder warning in place of file contents.\n- Binary document formats (PDF, DOC, DOCX, XLS, XLSX, PPT, PPTX) are not supported and will return an error.`,
+      description: "Read a file from the local filesystem.",
       parameters: {
         type: "object",
         properties: {