npm - @votadev/tooncode - Versions diffs - 2.2.3 → 2.2.6 - Mend

@votadev/tooncode 2.2.3 → 2.2.6

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (3) hide show

package/README.md CHANGED Viewed

@@ -1,243 +1,243 @@
-# ToonCode — Free AI Coding Agent CLI
-> **by VotaLab** | v2.0.0 | Claude Code alternative powered by free AI models
-```
-  ████████╗ ██████╗  ██████╗ ███╗   ██╗ ██████╗ ██████╗ ██████╗ ███████╗
-     ██║   ██║   ██║██║   ██║██╔██╗ ██║██║     ██║   ██║██║  ██║█████╗
-     ╚═╝    ╚═════╝  ╚═════╝ ╚═╝  ╚═══╝ ╚═════╝ ╚═════╝ ╚═════╝ ╚══════╝
-```
-ToonCode is a **free, open-source** AI coding agent that runs in your terminal. It uses free AI models (no API key needed) and provides a Claude Code-like experience with 20+ tools, multi-agent teams, browser automation, and more.
----
-## Requirements
-- **Python 3.10+**
-- **Git**
-- **Node.js 16+** (optional, for npm install method)
----
-## Installation
-### Method 1: One-line install (recommended)
-**Mac / Linux:**
-```bash
-curl -fsSL https://raw.githubusercontent.com/votalab/tooncode/main/install.sh | bash
-```
-**Windows (PowerShell):**
-```powershell
-irm https://raw.githubusercontent.com/votalab/tooncode/main/install.bat -OutFile install.bat; .\install.bat
-```
-### Method 2: Git clone (manual)
-```bash
-# Clone
-git clone https://github.com/votalab/tooncode.git ~/.tooncode
-# Install dependencies
-pip install -r ~/.tooncode/requirements.txt
-# Optional: Playwright for browser tool
-pip install playwright && playwright install chromium
-# Run
-python ~/.tooncode/tooncode.py
-```
-**Add to PATH (run from anywhere):**
-Mac/Linux — add to `~/.bashrc` or `~/.zshrc`:
-```bash
-alias tooncode="python ~/.tooncode/tooncode.py"
-```
-Windows — create `tooncode.cmd` in a PATH directory:
-```cmd
-@echo off
-python %USERPROFILE%\.tooncode\tooncode.py %*
-```
-### Method 3: npm
-```bash
-npm install -g tooncode
-```
-### Method 4: Copy the folder
-Just copy the entire `tooncode` folder anywhere and run:
-```bash
-cd /path/to/tooncode
-pip install -r requirements.txt
-python tooncode.py
-```
----
-## Quick Start
-```bash
-tooncode                          # Start in current directory
-tooncode big-pickle              # Start with specific model
-tooncode minimax-m2.5-free ~/my-project  # Model + directory
-```
----
-## Features
-### 20 Tools
-| Tool | Description |
-|------|-------------|
-| `bash` | Run shell commands (auto-detects PowerShell/bash) |
-| `read` | Read files with line numbers |
-| `write` | Create/overwrite files (shows diff) |
-| `edit` | String replacement in files (shows diff) |
-| `multi_edit` | Batch edits in one call |
-| `glob` | Find files by pattern |
-| `grep` | Search file contents with regex |
-| `list_dir` | Browse directory structure |
-| `web_search` | Search the web (DuckDuckGo) |
-| `web_fetch` | Fetch web page content |
-| `http` | Full HTTP client (GET/POST/PUT/DELETE + auth) |
-| `browser` | Playwright browser: open, click, fill, screenshot, console, network |
-| `screenshot` | Capture desktop/window screenshot |
-| `spawn_agent` | Spawn sub-agent (coder/reviewer/tester/researcher) |
-| `memory_save` | Save notes for future sessions |
-| `memory_search` | Search past memories |
-| `task_create/update/list` | Task management |
-| `bosshelp` | Escalate to Claude Code or fallback model |
-### Slash Commands
-| Command | Description |
-|---------|-------------|
-| `/help` | Show all commands |
-| `/model [name]` | Switch AI model |
-| `/boss <task>` | Create task plan (uses Claude Code or own model) |
-| `/team <task>` | Multi-agent team: planner + frontend + backend + reviewer + tester |
-| `/plan` | Toggle Plan Mode (read-only) |
-| `/do` | Execute pending tasks |
-| `/tasks` | Show task progress |
-| `/continue`, `/c` | Continue where AI left off |
-| `/save`, `/s` | Save session |
-| `/resume`, `/r` | Load saved session |
-| `/compact` | Compress context |
-| `/memory` | Browse saved memories |
-| `/paste`, `/v` | Send clipboard content |
-| `/send <msg>` | Chat between ToonCode windows |
-| `/bg`, `/ps` | Manage background processes |
-| `/commit` | Git add & commit |
-| `/diff` | Git diff |
-| `/status` | Git status |
-| `/undo` | Undo last file edit |
-| `/init` | Create TOONCODE.md (auto-imports CLAUDE.md etc.) |
-| `/config` | Show/edit config |
-| `/skills` | List 32 built-in skills |
-| `/cost` | Token usage stats |
-### Multi-Agent Team
-```bash
-/team สร้างเว็บ portfolio           # Default: planner + frontend + backend
-/team สร้าง API --roles planner,backend,tester   # Custom roles
-```
-Agents coordinate automatically via shared channel:
-- 📋 **planner** — Creates plan, assigns work
-- 🎨 **frontend** — Builds UI (HTML/CSS/JS)
-- ⚙️ **backend** — Builds API/server logic
-- 🔍 **reviewer** — Reviews code, fixes bugs
-- 🧪 **tester** — Writes & runs tests
-### 32 Built-in Skills
-```bash
-/refactor src/utils.py    # Refactor code
-/review                    # Code review
-/test api/users.js         # Generate tests
-/fix                       # Fix bugs
-/explain server.py         # Explain code
-/optimize                  # Performance optimization
-/docker                    # Create Dockerfile
-/commit                    # Smart git commit
-/audit                     # Security audit
-```
-### Available Models
-| Model | Context | Notes |
-|-------|---------|-------|
-| `minimax-m2.5-free` | 200K | Default, good all-round |
-| `big-pickle` | 200K | Strong coding |
-| `nemotron-3-super-free` | 131K | Fast |
-| `gpt-5-nano` | 1M | Largest context |
-Auto-fallback: if a model fails, ToonCode automatically tries the next one.
----
-## Project Context
-ToonCode auto-reads these files from your project:
-- `TOONCODE.md` — ToonCode project context
-- `CLAUDE.md` — Claude Code context (auto-imported)
-- `GEMINI.md` — Gemini context
-- `.cursorrules` — Cursor rules
-- `COPILOT.md` — Copilot instructions
-Run `/init` to auto-create `TOONCODE.md` with imported content.
----
-## Browser Tool
-Requires Playwright:
-```bash
-pip install playwright && playwright install chromium
-```
-AI can control a real browser:
-```
-❯ เปิด google ค้นหา "ToonCode" แล้วสรุปผลลัพธ์
-AI → browser open https://google.com/search?q=ToonCode
-AI → browser text  (reads search results)
-AI → browser console  (checks JS errors)
-AI → browser screenshot  (captures page)
-```
----
-## Update
-```bash
-cd ~/.tooncode && git pull
-```
----
-## Cross-Platform
-| Feature | Windows | Mac | Linux |
-|---------|:-------:|:---:|:-----:|
-| All tools | ✅ | ✅ | ✅ |
-| Shell | PowerShell | bash | bash |
-| Screenshot | ImageGrab | screencapture | scrot |
-| Clipboard | Get-Clipboard | pbpaste | xclip |
-| Browser | ✅ | ✅ | ✅ |
----
-## License
-MIT — Free to use, modify, and distribute.
-**Made with ❤️ by VotaLab**
+# ToonCode — 🇹🇭 Thai Coding Agent CLI
+> **by VotaLab** | v2.0.0 | Claude Code alternative powered by free AI models
+```
+  ████████╗ ██████╗  ██████╗ ███╗   ██╗ ██████╗ ██████╗ ██████╗ ███████╗
+     ██║   ██║   ██║██║   ██║██╔██╗ ██║██║     ██║   ██║██║  ██║█████╗
+     ╚═╝    ╚═════╝  ╚═════╝ ╚═╝  ╚═══╝ ╚═════╝ ╚═════╝ ╚═════╝ ╚══════╝
+```
+ToonCode is a **free, open-source** AI coding agent that runs in your terminal. It uses free AI models (no API key needed) and provides a Claude Code-like experience with 20+ tools, multi-agent teams, browser automation, and more.
+---
+## Requirements
+- **Python 3.10+**
+- **Git**
+- **Node.js 16+** (optional, for npm install method)
+---
+## Installation
+### Method 1: One-line install (recommended)
+**Mac / Linux:**
+```bash
+curl -fsSL https://raw.githubusercontent.com/votalab/tooncode/main/install.sh | bash
+```
+**Windows (PowerShell):**
+```powershell
+irm https://raw.githubusercontent.com/votalab/tooncode/main/install.bat -OutFile install.bat; .\install.bat
+```
+### Method 2: Git clone (manual)
+```bash
+# Clone
+git clone https://github.com/votalab/tooncode.git ~/.tooncode
+# Install dependencies
+pip install -r ~/.tooncode/requirements.txt
+# Optional: Playwright for browser tool
+pip install playwright && playwright install chromium
+# Run
+python ~/.tooncode/tooncode.py
+```
+**Add to PATH (run from anywhere):**
+Mac/Linux — add to `~/.bashrc` or `~/.zshrc`:
+```bash
+alias tooncode="python ~/.tooncode/tooncode.py"
+```
+Windows — create `tooncode.cmd` in a PATH directory:
+```cmd
+@echo off
+python %USERPROFILE%\.tooncode\tooncode.py %*
+```
+### Method 3: npm
+```bash
+npm install -g tooncode
+```
+### Method 4: Copy the folder
+Just copy the entire `tooncode` folder anywhere and run:
+```bash
+cd /path/to/tooncode
+pip install -r requirements.txt
+python tooncode.py
+```
+---
+## Quick Start
+```bash
+tooncode                          # Start in current directory
+tooncode big-pickle              # Start with specific model
+tooncode minimax-m2.5-free ~/my-project  # Model + directory
+```
+---
+## Features
+### 20 Tools
+| Tool | Description |
+|------|-------------|
+| `bash` | Run shell commands (auto-detects PowerShell/bash) |
+| `read` | Read files with line numbers |
+| `write` | Create/overwrite files (shows diff) |
+| `edit` | String replacement in files (shows diff) |
+| `multi_edit` | Batch edits in one call |
+| `glob` | Find files by pattern |
+| `grep` | Search file contents with regex |
+| `list_dir` | Browse directory structure |
+| `web_search` | Search the web (DuckDuckGo) |
+| `web_fetch` | Fetch web page content |
+| `http` | Full HTTP client (GET/POST/PUT/DELETE + auth) |
+| `browser` | Playwright browser: open, click, fill, screenshot, console, network |
+| `screenshot` | Capture desktop/window screenshot |
+| `spawn_agent` | Spawn sub-agent (coder/reviewer/tester/researcher) |
+| `memory_save` | Save notes for future sessions |
+| `memory_search` | Search past memories |
+| `task_create/update/list` | Task management |
+| `bosshelp` | Escalate to Claude Code or fallback model |
+### Slash Commands
+| Command | Description |
+|---------|-------------|
+| `/help` | Show all commands |
+| `/model [name]` | Switch AI model |
+| `/boss <task>` | Create task plan (uses Claude Code or own model) |
+| `/team <task>` | Multi-agent team: planner + frontend + backend + reviewer + tester |
+| `/plan` | Toggle Plan Mode (read-only) |
+| `/do` | Execute pending tasks |
+| `/tasks` | Show task progress |
+| `/continue`, `/c` | Continue where AI left off |
+| `/save`, `/s` | Save session |
+| `/resume`, `/r` | Load saved session |
+| `/compact` | Compress context |
+| `/memory` | Browse saved memories |
+| `/paste`, `/v` | Send clipboard content |
+| `/send <msg>` | Chat between ToonCode windows |
+| `/bg`, `/ps` | Manage background processes |
+| `/commit` | Git add & commit |
+| `/diff` | Git diff |
+| `/status` | Git status |
+| `/undo` | Undo last file edit |
+| `/init` | Create TOONCODE.md (auto-imports CLAUDE.md etc.) |
+| `/config` | Show/edit config |
+| `/skills` | List 32 built-in skills |
+| `/cost` | Token usage stats |
+### Multi-Agent Team
+```bash
+/team สร้างเว็บ portfolio           # Default: planner + frontend + backend
+/team สร้าง API --roles planner,backend,tester   # Custom roles
+```
+Agents coordinate automatically via shared channel:
+- 📋 **planner** — Creates plan, assigns work
+- 🎨 **frontend** — Builds UI (HTML/CSS/JS)
+- ⚙️ **backend** — Builds API/server logic
+- 🔍 **reviewer** — Reviews code, fixes bugs
+- 🧪 **tester** — Writes & runs tests
+### 32 Built-in Skills
+```bash
+/refactor src/utils.py    # Refactor code
+/review                    # Code review
+/test api/users.js         # Generate tests
+/fix                       # Fix bugs
+/explain server.py         # Explain code
+/optimize                  # Performance optimization
+/docker                    # Create Dockerfile
+/commit                    # Smart git commit
+/audit                     # Security audit
+```
+### Available Models
+| Model | Context | Notes |
+|-------|---------|-------|
+| `minimax-m2.5-free` | 200K | Default, good all-round |
+| `big-pickle` | 200K | Strong coding |
+| `nemotron-3-super-free` | 131K | Fast |
+| `gpt-5-nano` | 1M | Largest context |
+Auto-fallback: if a model fails, ToonCode automatically tries the next one.
+---
+## Project Context
+ToonCode auto-reads these files from your project:
+- `TOONCODE.md` — ToonCode project context
+- `CLAUDE.md` — Claude Code context (auto-imported)
+- `GEMINI.md` — Gemini context
+- `.cursorrules` — Cursor rules
+- `COPILOT.md` — Copilot instructions
+Run `/init` to auto-create `TOONCODE.md` with imported content.
+---
+## Browser Tool
+Requires Playwright:
+```bash
+pip install playwright && playwright install chromium
+```
+AI can control a real browser:
+```
+❯ เปิด google ค้นหา "ToonCode" แล้วสรุปผลลัพธ์
+AI → browser open https://google.com/search?q=ToonCode
+AI → browser text  (reads search results)
+AI → browser console  (checks JS errors)
+AI → browser screenshot  (captures page)
+```
+---
+## Update
+```bash
+cd ~/.tooncode && git pull
+```
+---
+## Cross-Platform
+| Feature | Windows | Mac | Linux |
+|---------|:-------:|:---:|:-----:|
+| All tools | ✅ | ✅ | ✅ |
+| Shell | PowerShell | bash | bash |
+| Screenshot | ImageGrab | screencapture | scrot |
+| Clipboard | Get-Clipboard | pbpaste | xclip |
+| Browser | ✅ | ✅ | ✅ |
+---
+## License
+MIT — Free to use, modify, and distribute.
+**Made with ❤️ by VotaLab**

package/package.json CHANGED Viewed

@@ -1,42 +1,42 @@
-{
-  "name": "@votadev/tooncode",
-  "version": "2.2.3",
-  "description": "Free AI Coding Agent CLI — Claude Code alternative powered by free models. 20 tools, multi-agent team, browser automation, MCP servers.",
-  "author": "VotaLab",
-  "license": "MIT",
-  "keywords": [
-    "ai",
-    "coding",
-    "agent",
-    "cli",
-    "claude",
-    "free",
-    "tooncode",
-    "copilot",
-    "cursor"
-  ],
-  "repository": {
-    "type": "git",
-    "url": "git+https://github.com/votalab/tooncode.git"
-  },
-  "homepage": "https://github.com/votalab/tooncode#readme",
-  "bin": {
-    "tooncode": "bin/tooncode.js"
-  },
-  "files": [
-    "bin/",
-    "tooncode.py",
-    "__main__.py",
-    "skills/",
-    "requirements.txt",
-    "setup.py",
-    "tooncode.cmd",
-    "README.md"
-  ],
-  "scripts": {
-    "postinstall": "node bin/postinstall.js"
-  },
-  "engines": {
-    "node": ">=16"
-  }
-}
+{
+  "name": "@votadev/tooncode",
+  "version": "2.2.6",
+  "description": "🇹🇭 Thai Coding Agent CLI — Claude Code alternative powered by free models. 20 tools, multi-agent team, browser automation, MCP servers.",
+  "author": "VotaLab",
+  "license": "MIT",
+  "keywords": [
+    "ai",
+    "coding",
+    "agent",
+    "cli",
+    "claude",
+    "free",
+    "tooncode",
+    "copilot",
+    "cursor"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/votalab/tooncode.git"
+  },
+  "homepage": "https://github.com/votalab/tooncode#readme",
+  "bin": {
+    "tooncode": "bin/tooncode.js"
+  },
+  "files": [
+    "bin/",
+    "tooncode.py",
+    "__main__.py",
+    "skills/",
+    "requirements.txt",
+    "setup.py",
+    "tooncode.cmd",
+    "README.md"
+  ],
+  "scripts": {
+    "postinstall": "node bin/postinstall.js"
+  },
+  "engines": {
+    "node": ">=16"
+  }
+}

package/tooncode.py CHANGED Viewed

@@ -8,7 +8,7 @@ Usage:
     python tooncode.py
 """
-VERSION = "2.2.3"
+VERSION = "2.2.6"
 import httpx
 import json
@@ -2915,10 +2915,40 @@ def _get_current_api_url() -> str:
     return _get_api_url()
+def _validate_messages(messages: list):
+    """Fix message ordering issues before sending to API."""
+    # Rule: tool_result must follow assistant with tool_use
+    # Rule: alternating user/assistant (with tool_result counting as user)
+    fixed = []
+    for i, msg in enumerate(messages):
+        role = msg.get("role", "")
+        # Skip empty messages
+        content = msg.get("content", [])
+        if not content:
+            continue
+        # Prevent consecutive same-role messages (except tool_result after user)
+        if fixed:
+            prev_role = fixed[-1].get("role", "")
+            if role == prev_role and role == "user":
+                # Check if this is tool_result (allowed after tool_use response)
+                is_tool_result = isinstance(content, list) and any(
+                    b.get("type") == "tool_result" for b in content if isinstance(b, dict))
+                if not is_tool_result:
+                    fixed.append({"role": "assistant", "content": [{"type": "text", "text": "(continuing)"}]})
+            elif role == prev_role and role == "assistant":
+                fixed.append({"role": "user", "content": [{"type": "text", "text": "(continue)"}]})
+        fixed.append(msg)
+    messages.clear()
+    messages.extend(fixed)
 def stream_response(messages: list, renderer: StreamRenderer) -> dict:
     """Send request and stream the response, returning parsed content blocks."""
     global total_input_tokens, total_output_tokens, last_input_tokens
+    # Validate message ordering
+    _validate_messages(messages)
     system_prompt = build_system_prompt()
     body = {
         "model": MODEL,
@@ -3285,7 +3315,7 @@ Be specific. Give copy-paste ready solutions."""
         console.print("[bold magenta][bosshelp] Asking Claude Code...[/bold magenta]")
         try:
             result = subprocess.run(
-                [claude_cmd, "--print"],
+                [claude_cmd, "--print", "--dangerously-skip-permissions"],
                 input=help_prompt,
                 capture_output=True, text=True, cwd=CWD, timeout=180,
                 encoding="utf-8", errors="replace",
@@ -4197,7 +4227,7 @@ OUTPUT THE JSON ARRAY NOW:"""
                 console.print("[dim]Using Claude Code CLI...[/dim]")
                 try:
                     result = subprocess.run(
-                        [claude_cmd, "--print"],
+                        [claude_cmd, "--print", "--dangerously-skip-permissions"],
                         input=boss_prompt,
                         capture_output=True, text=True, cwd=CWD, timeout=180,
                         encoding="utf-8", errors="replace",
@@ -4696,7 +4726,7 @@ def print_banner():
     tagline.append("  |  ", style="dim")
     tagline.append(f"v{VERSION}", style="bold #00ccff")
     tagline.append("  |  ", style="dim")
-    tagline.append("Free AI Coding Agent", style="dim white")
+    tagline.append("🇹🇭 Thai Coding Agent", style="dim white")
     # Count stats
     install_dir = os.path.dirname(os.path.abspath(__file__))
@@ -5312,21 +5342,21 @@ def main(_initial_prompt=None):
                                     )
                                     _last_errors.clear()
                                     _post_bosshelp = True
-                                    # Inject boss answer as a separate user message after tool results
+                                    # MUST append tool_results first (API requires it after tool_use)
                                     messages.append({"role": "user", "content": tool_results})
-                                    messages.append({"role": "assistant", "content": [{"type": "text", "text": "I'm stuck. Let me check the boss's advice."}]})
-                                    # Parse boss answer for file paths and code blocks
-                                    fix_instructions = f"[BOSSHELP - APPLY NOW]\n{boss_answer}\n\n"
-                                    fix_instructions += "STEP-BY-STEP INSTRUCTIONS:\n"
-                                    fix_instructions += "1. Read each file mentioned above using the read tool\n"
-                                    fix_instructions += "2. Apply each code change using the edit tool (oldString → newString)\n"
-                                    fix_instructions += "3. If the fix says to run a command, use the bash tool\n"
-                                    fix_instructions += "4. After applying, verify with bash (e.g. python -c 'import module')\n"
-                                    fix_instructions += "\nDo NOT explain. Do NOT repeat failed approach. USE TOOLS NOW."
+                                    # Then add bosshelp as assistant+user pair
+                                    fix_instructions = f"[BOSSHELP]\n{boss_answer}\n\n"
+                                    fix_instructions += "APPLY THIS FIX NOW:\n"
+                                    fix_instructions += "1. Read each file mentioned\n"
+                                    fix_instructions += "2. Use edit tool to apply changes\n"
+                                    fix_instructions += "3. Run commands if needed\n"
+                                    fix_instructions += "4. Verify the fix works\n"
+                                    fix_instructions += "USE TOOLS. Do NOT explain."
+                                    messages.append({"role": "assistant", "content": [{"type": "text", "text": "Let me apply the fix."}]})
                                     messages.append({"role": "user", "content": [{"type": "text", "text":
                                         fix_instructions,
                                         "cache_control": {"type": "ephemeral"}}]})
-                                    continue  # Skip the normal append below
+                                    continue
                         else:
                             # Successful tool = reset error tracking
                             _last_errors.clear()
@@ -5400,7 +5430,7 @@ def _cli():
     parser = argparse.ArgumentParser(
         prog="tooncode",
-        description="ToonCode — Free AI Coding Agent CLI by VotaLab",
+        description="ToonCode — 🇹🇭 Thai Coding Agent CLI by VotaLab",
         epilog="""Examples:
   tooncode                     Start interactive mode
   tooncode -m big-pickle       Use specific model