feat: Add grep and websearch tools

Introduces two new tools: `grep` for filtering text/files by regex, and `websearch` for querying the web using LLM models. Updates package version.

Author: iovdin

README.md

@@ -13,8 +13,10 @@ Basic toolset for [Tune](https://github.com/iovdin/tune).
   - [sh](#sh) execute shell command
   - [cmd](#cmd) execute Windows cmd command
   - [powershell](#powershell) execute PowerShell command
+  - [grep](#grep) search for patterns in text or files
   - [osa](#osa) manage reminders/notes/calendar (AppleScript/macOS)
   - [jina_r](#jina_r) fetch webpage content
+  - [websearch](#websearch) search the web with web-enabled llms
   - [list](#list) keep list of tasks todo (loops for LLM)
   - [sqlite](#sqlite) execute sqlite queries 
   - [py](#py) run python code
@@ -175,6 +177,22 @@ TotalPhysicalMemory : 17179869184
 CsProcessors        : {Intel(R) Core(TM) i7-10700K CPU @ 3.80GHz}
 ```
 
+### `grep`
+Search for patterns in text or files using regular expressions
+```chat
+user: @grep
+find all lines containing "TODO" in myfile.js
+tool_call: grep {"filename":"myfile.js","regex":"TODO"}
+tool_result: 
+// TODO: refactor this function
+// TODO: add error handling
+
+system: 
+TODOS:
+@{ myfile.js | proc grep regex=TODO }
+
+```
+
 ### `osa`
 AppleScript tool, manage reminders, notes, calendar etc on osx
 ```chat
@@ -219,7 +237,25 @@ Tune is a versatile toolkit designed for developers and users to effectively int
 <cut for brevity>
 ```
 
+### `websearch`
+Search the web with web enabled llms
+Supports search with  `perplexity/sonar`, `perplexity/sonar-pro`, `gpt-4o-search-preview`, `gpt-4o-mini-search-preview` models via the `model` parameter (defaults to `perplexity/sonar`).
+
+```chat
+user: @websearch
+latest ai news
+
+assistant:
+
+tool_call: websearch {"model":"perplexity/sonar"}
+latest AI news
+tool_result:
+The latest AI news in October 2025 highlights significant investments, new projects, policy developments, and advances across various sectors:
+
+- Major companies including Microsoft, Google, Nvidia, OpenAI, Salesforce, and CoreWeave have pledged over £31 billion in capital expenditure focused on AI data centers and infrastructure upgrades[1].
+```
 
+The websearch tool provides up-to-date information by querying the web through AI-powered search models. You can specify different Perplexity models like `perplexity/sonar-pro` for more advanced searches.
 
 ### `list`
 Keep list of tasks to do

package.json

@@ -1,6 +1,6 @@
 {
   "name": "tune-basic-toolset",
-  "version": "0.1.10",
+  "version": "0.1.11",
   "description": "Basic toolset for tune",
   "main": "src/index.js",
   "files": [

src/grep.schema.json

@@ -0,0 +1,25 @@
+{
+  "description": "filter lines with regex",
+  "parameters": {
+    "type": "object",
+    "properties": {
+      "text": {
+        "type": "string",
+        "description": "Text to be filtered"
+      },
+      "filename": {
+        "type": "string",
+        "description": "if text is not set, filename will be read and filtered"
+      },
+      "regex": {
+        "type": "string",
+        "description": "js regex to filter lines in file or text. passed as first arguments to new RegExp "
+      },
+      "regex_flags": {
+        "type": "string",
+        "description": "flags to be passed to new RegExp constructor"
+      }
+    },
+    "required": ["regex"]
+  }
+}

src/grep.tool.js

@@ -0,0 +1,16 @@
+module.exports = async function grep({filename, text, regex, regex_flags}, ctx) {
+  if (!text && filename) {
+    const n = await ctx.resolve(filename)
+    if (!n) 
+      return `${filename} not found`
+    
+    text = await n.read()
+  }
+
+  if (!text) {
+    return "content is empty"
+  }
+
+  const r = new RegExp(regex, regex_flags)
+  return text.split(/\r?\n/).filter(line => r.test(line)).join("\n").replaceAll("@", "\\@")
+}

src/websearch.schema.json

@@ -0,0 +1,23 @@
+{
+  "description": "Does a websearch using llm",
+  "parameters": {
+    "type": "object",
+    "properties": {
+      "text": {
+        "type": "string",
+        "description": "web search query"
+      },
+      "model": {
+        "type": "string",
+        "enum": [
+          "perplexity/sonar", 
+          "perplexity/sonar-pro", 
+          "gpt-4o-search-preview", 
+          "gpt-4o-mini-search-preview"
+        ],
+        "description": "model to do websearch, default is perplexity/sonar"
+      }
+    },
+    "required": ["text"]
+  }
+}

src/websearch.tool.chat

@@ -0,0 +1,3 @@
+user:
+@{ model | init perplexity/sonar | resolve }
+@text

test/p05.txt

@@ -0,0 +1,17 @@
+a
+b
+c
+-------------------------------
+<<<<<<< ORIGINAL
+a
+b
+c
+=======
+d
+e
+f
+>>>>>>> UPDATED
+-------------------------------
+d
+e
+f

test/p06.txt

@@ -0,0 +1,19 @@
+x
+  (aif user.mangopayId
+    (a (href: (+ "https://dashboard.mangopay.com/User/" it "/Details") target:"_blank")
+      "Mangopay wallet"))
+y
+-------------------------------
+<<<<<<< ORIGINAL
+  (aif user.mangopayId
+    (a (href: (+ "https://dashboard.mangopay.com/User/" it "/Details") target:"_blank")
+      "Mangopay wallet"))
+=======
+  (aif user.mangopayId
+    (span "Mangopay: " it))
+>>>>>> UPDATED
+-------------------------------
+x
+  (aif user.mangopayId
+    (span "Mangopay: " it))
+y

test/p07.txt

@@ -0,0 +1,13 @@
+foo
+bar
+baz
+-------------------------------
+<<<<<<<< ORIGINAL
+bar
+=======
+BAR
+>>>>>>>> UPDATED
+-------------------------------
+foo
+BAR
+baz

test/p08.txt

@@ -0,0 +1,13 @@
+a
+b
+c
+-------------------------------
+<<<<<<< ORIGINAL
+b
+=======
+B
+>>>>>> UPDATED
+-------------------------------
+a
+B
+c